Merge remote-tracking branch 'upstream/master' into multitalk

README.md

@@ -206,14 +206,32 @@ Put your SD checkpoints (the huge ckpt/safetensors files) in: models/checkpoints
 Put your VAE in: models/vae
 
 
-### AMD GPUs (Linux only)
+### AMD GPUs (Linux)
+
 AMD users can install rocm and pytorch with pip if you don't have it already installed, this is the command to install the stable version:
 
 ```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.4```
 
-This is the command to install the nightly with ROCm 6.4 which might have some performance improvements:
+This is the command to install the nightly with ROCm 7.0 which might have some performance improvements:
 
-```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.4```
+```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm7.0```
+
+
+### AMD GPUs (Experimental: Windows and Linux), RDNA 3, 3.5 and 4 only.
+
+These have less hardware support than the builds above but they work on windows. You also need to install the pytorch version specific to your hardware.
+
+RDNA 3 (RX 7000 series):
+
+```pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx110X-dgpu/```
+
+RDNA 3.5 (Strix halo/Ryzen AI Max+ 365):
+
+```pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx1151/```
+
+RDNA 4 (RX 9000 series):
+
+```pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx120X-all/```
 
 ### Intel GPUs (Windows and Linux)
 
@@ -270,12 +288,6 @@ You can install ComfyUI in Apple Mac silicon (M1 or M2) with any recent macOS ve
 
 > **Note**: Remember to add your models, VAE, LoRAs etc. to the corresponding Comfy folders, as discussed in [ComfyUI manual installation](#manual-install-windows-linux).
 
-#### DirectML (AMD Cards on Windows)
-
-This is very badly supported and is not recommended. There are some unofficial builds of pytorch ROCm on windows that exist that will give you a much better experience than this. This readme will be updated once official pytorch ROCm builds for windows come out.
-
-```pip install torch-directml``` Then you can launch ComfyUI with: ```python main.py --directml```
-
 #### Ascend NPUs
 
 For models compatible with Ascend Extension for PyTorch (torch_npu). To get started, ensure your environment meets the prerequisites outlined on the [installation](https://ascend.github.io/docs/sources/ascend/quick_install.html) page. Here's a step-by-step guide tailored to your platform and installation method:

comfy/ldm/ace/vae/music_dcae_pipeline.py

@@ -23,8 +23,6 @@ class MusicDCAE(torch.nn.Module):
         else:
             self.source_sample_rate = source_sample_rate
 
-        # self.resampler = torchaudio.transforms.Resample(source_sample_rate, 44100)
-
         self.transform = transforms.Compose([
             transforms.Normalize(0.5, 0.5),
         ])
@@ -37,10 +35,6 @@ class MusicDCAE(torch.nn.Module):
         self.scale_factor = 0.1786
         self.shift_factor = -1.9091
 
-    def load_audio(self, audio_path):
-        audio, sr = torchaudio.load(audio_path)
-        return audio, sr
-
     def forward_mel(self, audios):
         mels = []
         for i in range(len(audios)):
@@ -73,10 +67,8 @@ class MusicDCAE(torch.nn.Module):
             latent = self.dcae.encoder(mel.unsqueeze(0))
             latents.append(latent)
         latents = torch.cat(latents, dim=0)
-        # latent_lengths = (audio_lengths / sr * 44100 / 512 / self.time_dimention_multiple).long()
         latents = (latents - self.shift_factor) * self.scale_factor
         return latents
-        # return latents, latent_lengths
 
     @torch.no_grad()
     def decode(self, latents, audio_lengths=None, sr=None):
@@ -91,9 +83,7 @@ class MusicDCAE(torch.nn.Module):
             wav = self.vocoder.decode(mels[0]).squeeze(1)
 
             if sr is not None:
-                # resampler = torchaudio.transforms.Resample(44100, sr).to(latents.device).to(latents.dtype)
                 wav = torchaudio.functional.resample(wav, 44100, sr)
-                # wav = resampler(wav)
             else:
                 sr = 44100
             pred_wavs.append(wav)
@@ -101,7 +91,6 @@ class MusicDCAE(torch.nn.Module):
         if audio_lengths is not None:
             pred_wavs = [wav[:, :length].cpu() for wav, length in zip(pred_wavs, audio_lengths)]
         return torch.stack(pred_wavs)
-        # return sr, pred_wavs
 
     def forward(self, audios, audio_lengths=None, sr=None):
         latents, latent_lengths = self.encode(audios=audios, audio_lengths=audio_lengths, sr=sr)

comfy/ldm/wan/model.py

@@ -915,7 +915,7 @@ class MotionEncoder_tc(nn.Module):
     def __init__(self,
                  in_dim: int,
                  hidden_dim: int,
-                 num_heads=int,
+                 num_heads: int,
                  need_global=True,
                  dtype=None,
                  device=None,

comfy/model_detection.py

@@ -365,8 +365,8 @@ def detect_unet_config(state_dict, key_prefix, metadata=None):
         dit_config["patch_size"] = 2
         dit_config["in_channels"] = 16
         dit_config["dim"] = 2304
-        dit_config["cap_feat_dim"] = 2304
-        dit_config["n_layers"] = 26
+        dit_config["cap_feat_dim"] = state_dict['{}cap_embedder.1.weight'.format(key_prefix)].shape[1]
+        dit_config["n_layers"] = count_blocks(state_dict_keys, '{}layers.'.format(key_prefix) + '{}.')
         dit_config["n_heads"] = 24
         dit_config["n_kv_heads"] = 8
         dit_config["qk_norm"] = True

comfy/model_patcher.py

@@ -123,16 +123,30 @@ def move_weight_functions(m, device):
     return memory
 
 class LowVramPatch:
-    def __init__(self, key, patches):
+    def __init__(self, key, patches, convert_func=None, set_func=None):
         self.key = key
         self.patches = patches
+        self.convert_func = convert_func
+        self.set_func = set_func
+
     def __call__(self, weight):
         intermediate_dtype = weight.dtype
+        if self.convert_func is not None:
+            weight = self.convert_func(weight.to(dtype=torch.float32, copy=True), inplace=True)
+
         if intermediate_dtype not in [torch.float32, torch.float16, torch.bfloat16]: #intermediate_dtype has to be one that is supported in math ops
             intermediate_dtype = torch.float32
-            return comfy.float.stochastic_rounding(comfy.lora.calculate_weight(self.patches[self.key], weight.to(intermediate_dtype), self.key, intermediate_dtype=intermediate_dtype), weight.dtype, seed=string_to_seed(self.key))
+            out = comfy.lora.calculate_weight(self.patches[self.key], weight.to(intermediate_dtype), self.key, intermediate_dtype=intermediate_dtype)
+            if self.set_func is None:
+                return comfy.float.stochastic_rounding(out, weight.dtype, seed=string_to_seed(self.key))
+            else:
+                return self.set_func(out, seed=string_to_seed(self.key), return_weight=True)
 
-        return comfy.lora.calculate_weight(self.patches[self.key], weight, self.key, intermediate_dtype=intermediate_dtype)
+        out = comfy.lora.calculate_weight(self.patches[self.key], weight, self.key, intermediate_dtype=intermediate_dtype)
+        if self.set_func is not None:
+            return self.set_func(out, seed=string_to_seed(self.key), return_weight=True).to(dtype=intermediate_dtype)
+        else:
+            return out
 
 def get_key_weight(model, key):
     set_func = None
@@ -657,13 +671,15 @@ class ModelPatcher:
                         if force_patch_weights:
                             self.patch_weight_to_device(weight_key)
                         else:
-                            m.weight_function = [LowVramPatch(weight_key, self.patches)]
+                            _, set_func, convert_func = get_key_weight(self.model, weight_key)
+                            m.weight_function = [LowVramPatch(weight_key, self.patches, convert_func, set_func)]
                             patch_counter += 1
                     if bias_key in self.patches:
                         if force_patch_weights:
                             self.patch_weight_to_device(bias_key)
                         else:
-                            m.bias_function = [LowVramPatch(bias_key, self.patches)]
+                            _, set_func, convert_func = get_key_weight(self.model, bias_key)
+                            m.bias_function = [LowVramPatch(bias_key, self.patches, convert_func, set_func)]
                             patch_counter += 1
 
                     cast_weight = True
@@ -825,10 +841,12 @@ class ModelPatcher:
                         module_mem += move_weight_functions(m, device_to)
                         if lowvram_possible:
                             if weight_key in self.patches:
-                                m.weight_function.append(LowVramPatch(weight_key, self.patches))
+                                _, set_func, convert_func = get_key_weight(self.model, weight_key)
+                                m.weight_function.append(LowVramPatch(weight_key, self.patches, convert_func, set_func))
                                 patch_counter += 1
                             if bias_key in self.patches:
-                                m.bias_function.append(LowVramPatch(bias_key, self.patches))
+                                _, set_func, convert_func = get_key_weight(self.model, bias_key)
+                                m.bias_function.append(LowVramPatch(bias_key, self.patches, convert_func, set_func))
                                 patch_counter += 1
                             cast_weight = True
 

comfy/model_sampling.py

@@ -21,17 +21,23 @@ def rescale_zero_terminal_snr_sigmas(sigmas):
     alphas_bar[-1] = 4.8973451890853435e-08
     return ((1 - alphas_bar) / alphas_bar) ** 0.5
 
+def reshape_sigma(sigma, noise_dim):
+    if sigma.nelement() == 1:
+        return sigma.view(())
+    else:
+        return sigma.view(sigma.shape[:1] + (1,) * (noise_dim - 1))
+
 class EPS:
     def calculate_input(self, sigma, noise):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         return noise / (sigma ** 2 + self.sigma_data ** 2) ** 0.5
 
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input - model_output * sigma
 
     def noise_scaling(self, sigma, noise, latent_image, max_denoise=False):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         if max_denoise:
             noise = noise * torch.sqrt(1.0 + sigma ** 2.0)
         else:
@@ -45,12 +51,12 @@ class EPS:
 
 class V_PREDICTION(EPS):
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input * self.sigma_data ** 2 / (sigma ** 2 + self.sigma_data ** 2) - model_output * sigma * self.sigma_data / (sigma ** 2 + self.sigma_data ** 2) ** 0.5
 
 class EDM(V_PREDICTION):
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input * self.sigma_data ** 2 / (sigma ** 2 + self.sigma_data ** 2) + model_output * sigma * self.sigma_data / (sigma ** 2 + self.sigma_data ** 2) ** 0.5
 
 class CONST:
@@ -58,15 +64,15 @@ class CONST:
         return noise
 
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input - model_output * sigma
 
     def noise_scaling(self, sigma, noise, latent_image, max_denoise=False):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         return sigma * noise + (1.0 - sigma) * latent_image
 
     def inverse_noise_scaling(self, sigma, latent):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (latent.ndim - 1))
+        sigma = reshape_sigma(sigma, latent.ndim)
         return latent / (1.0 - sigma)
 
 class X0(EPS):
@@ -80,16 +86,16 @@ class IMG_TO_IMG(X0):
 class COSMOS_RFLOW:
     def calculate_input(self, sigma, noise):
         sigma = (sigma / (sigma + 1))
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         return noise * (1.0 - sigma)
 
     def calculate_denoised(self, sigma, model_output, model_input):
         sigma = (sigma / (sigma + 1))
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input * (1.0 - sigma) - model_output * sigma
 
     def noise_scaling(self, sigma, noise, latent_image, max_denoise=False):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         noise = noise * sigma
         noise += latent_image
         return noise

comfy/ops.py

@@ -416,8 +416,10 @@ def scaled_fp8_ops(fp8_matrix_mult=False, scale_input=False, override_dtype=None
                 else:
                     return weight * self.scale_weight.to(device=weight.device, dtype=weight.dtype)
 
-            def set_weight(self, weight, inplace_update=False, seed=None, **kwargs):
+            def set_weight(self, weight, inplace_update=False, seed=None, return_weight=False, **kwargs):
                 weight = comfy.float.stochastic_rounding(weight / self.scale_weight.to(device=weight.device, dtype=weight.dtype), self.weight.dtype, seed=seed)
+                if return_weight:
+                    return weight
                 if inplace_update:
                     self.weight.data.copy_(weight)
                 else:

comfy/sd.py

@@ -890,6 +890,7 @@ class TEModel(Enum):
     QWEN25_3B = 10
     QWEN25_7B = 11
     BYT5_SMALL_GLYPH = 12
+    GEMMA_3_4B = 13
 
 def detect_te_model(sd):
     if "text_model.encoder.layers.30.mlp.fc1.weight" in sd:
@@ -912,6 +913,8 @@ def detect_te_model(sd):
             return TEModel.BYT5_SMALL_GLYPH
         return TEModel.T5_BASE
     if 'model.layers.0.post_feedforward_layernorm.weight' in sd:
+        if 'model.layers.0.self_attn.q_norm.weight' in sd:
+            return TEModel.GEMMA_3_4B
         return TEModel.GEMMA_2_2B
     if 'model.layers.0.self_attn.k_proj.bias' in sd:
         weight = sd['model.layers.0.self_attn.k_proj.bias']
@@ -1016,6 +1019,10 @@ def load_text_encoder_state_dicts(state_dicts=[], embedding_directory=None, clip
             clip_target.clip = comfy.text_encoders.lumina2.te(**llama_detect(clip_data))
             clip_target.tokenizer = comfy.text_encoders.lumina2.LuminaTokenizer
             tokenizer_data["spiece_model"] = clip_data[0].get("spiece_model", None)
+        elif te_model == TEModel.GEMMA_3_4B:
+            clip_target.clip = comfy.text_encoders.lumina2.te(**llama_detect(clip_data), model_type="gemma3_4b")
+            clip_target.tokenizer = comfy.text_encoders.lumina2.NTokenizer
+            tokenizer_data["spiece_model"] = clip_data[0].get("spiece_model", None)
         elif te_model == TEModel.LLAMA3_8:
             clip_target.clip = comfy.text_encoders.hidream.hidream_clip(**llama_detect(clip_data),
                                                                         clip_l=False, clip_g=False, t5=False, llama=True, dtype_t5=None, t5xxl_scaled_fp8=None)

comfy/text_encoders/llama.py

@@ -3,6 +3,7 @@ import torch.nn as nn
 from dataclasses import dataclass
 from typing import Optional, Any
 import math
+import logging
 
 from comfy.ldm.modules.attention import optimized_attention_for_device
 import comfy.model_management
@@ -28,6 +29,9 @@ class Llama2Config:
     mlp_activation = "silu"
     qkv_bias = False
     rope_dims = None
+    q_norm = None
+    k_norm = None
+    rope_scale = None
 
 @dataclass
 class Qwen25_3BConfig:
@@ -46,6 +50,9 @@ class Qwen25_3BConfig:
     mlp_activation = "silu"
     qkv_bias = True
     rope_dims = None
+    q_norm = None
+    k_norm = None
+    rope_scale = None
 
 @dataclass
 class Qwen25_7BVLI_Config:
@@ -64,6 +71,9 @@ class Qwen25_7BVLI_Config:
     mlp_activation = "silu"
     qkv_bias = True
     rope_dims = [16, 24, 24]
+    q_norm = None
+    k_norm = None
+    rope_scale = None
 
 @dataclass
 class Gemma2_2B_Config:
@@ -82,6 +92,32 @@ class Gemma2_2B_Config:
     mlp_activation = "gelu_pytorch_tanh"
     qkv_bias = False
     rope_dims = None
+    q_norm = None
+    k_norm = None
+    sliding_attention = None
+    rope_scale = None
+
+@dataclass
+class Gemma3_4B_Config:
+    vocab_size: int = 262208
+    hidden_size: int = 2560
+    intermediate_size: int = 10240
+    num_hidden_layers: int = 34
+    num_attention_heads: int = 8
+    num_key_value_heads: int = 4
+    max_position_embeddings: int = 131072
+    rms_norm_eps: float = 1e-6
+    rope_theta = [10000.0, 1000000.0]
+    transformer_type: str = "gemma3"
+    head_dim = 256
+    rms_norm_add = True
+    mlp_activation = "gelu_pytorch_tanh"
+    qkv_bias = False
+    rope_dims = None
+    q_norm = "gemma3"
+    k_norm = "gemma3"
+    sliding_attention = [False, False, False, False, False, 1024]
+    rope_scale = [1.0, 8.0]
 
 class RMSNorm(nn.Module):
     def __init__(self, dim: int, eps: float = 1e-5, add=False, device=None, dtype=None):
@@ -106,25 +142,40 @@ def rotate_half(x):
     return torch.cat((-x2, x1), dim=-1)
 
 
-def precompute_freqs_cis(head_dim, position_ids, theta, rope_dims=None, device=None):
-    theta_numerator = torch.arange(0, head_dim, 2, device=device).float()
-    inv_freq = 1.0 / (theta ** (theta_numerator / head_dim))
+def precompute_freqs_cis(head_dim, position_ids, theta, rope_scale=None, rope_dims=None, device=None):
+    if not isinstance(theta, list):
+        theta = [theta]
 
-    inv_freq_expanded = inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
-    position_ids_expanded = position_ids[:, None, :].float()
-    freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
-    emb = torch.cat((freqs, freqs), dim=-1)
-    cos = emb.cos()
-    sin = emb.sin()
-    if rope_dims is not None and position_ids.shape[0] > 1:
-        mrope_section = rope_dims * 2
-        cos = torch.cat([m[i % 3] for i, m in enumerate(cos.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
-        sin = torch.cat([m[i % 3] for i, m in enumerate(sin.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
-    else:
-        cos = cos.unsqueeze(1)
-        sin = sin.unsqueeze(1)
+    out = []
+    for index, t in enumerate(theta):
+        theta_numerator = torch.arange(0, head_dim, 2, device=device).float()
+        inv_freq = 1.0 / (t ** (theta_numerator / head_dim))
 
-    return (cos, sin)
+        if rope_scale is not None:
+            if isinstance(rope_scale, list):
+                inv_freq /= rope_scale[index]
+            else:
+                inv_freq /= rope_scale
+
+        inv_freq_expanded = inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
+        position_ids_expanded = position_ids[:, None, :].float()
+        freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
+        emb = torch.cat((freqs, freqs), dim=-1)
+        cos = emb.cos()
+        sin = emb.sin()
+        if rope_dims is not None and position_ids.shape[0] > 1:
+            mrope_section = rope_dims * 2
+            cos = torch.cat([m[i % 3] for i, m in enumerate(cos.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
+            sin = torch.cat([m[i % 3] for i, m in enumerate(sin.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
+        else:
+            cos = cos.unsqueeze(1)
+            sin = sin.unsqueeze(1)
+        out.append((cos, sin))
+
+    if len(out) == 1:
+        return out[0]
+
+    return out
 
 
 def apply_rope(xq, xk, freqs_cis):
@@ -152,6 +203,14 @@ class Attention(nn.Module):
         self.v_proj = ops.Linear(config.hidden_size, self.num_kv_heads * self.head_dim, bias=config.qkv_bias, device=device, dtype=dtype)
         self.o_proj = ops.Linear(self.inner_size, config.hidden_size, bias=False, device=device, dtype=dtype)
 
+        self.q_norm = None
+        self.k_norm = None
+
+        if config.q_norm == "gemma3":
+            self.q_norm = RMSNorm(self.head_dim, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
+        if config.k_norm == "gemma3":
+            self.k_norm = RMSNorm(self.head_dim, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
+
     def forward(
         self,
         hidden_states: torch.Tensor,
@@ -168,6 +227,11 @@ class Attention(nn.Module):
         xk = xk.view(batch_size, seq_length, self.num_kv_heads, self.head_dim).transpose(1, 2)
         xv = xv.view(batch_size, seq_length, self.num_kv_heads, self.head_dim).transpose(1, 2)
 
+        if self.q_norm is not None:
+            xq = self.q_norm(xq)
+        if self.k_norm is not None:
+            xk = self.k_norm(xk)
+
         xq, xk = apply_rope(xq, xk, freqs_cis=freqs_cis)
 
         xk = xk.repeat_interleave(self.num_heads // self.num_kv_heads, dim=1)
@@ -192,7 +256,7 @@ class MLP(nn.Module):
         return self.down_proj(self.activation(self.gate_proj(x)) * self.up_proj(x))
 
 class TransformerBlock(nn.Module):
-    def __init__(self, config: Llama2Config, device=None, dtype=None, ops: Any = None):
+    def __init__(self, config: Llama2Config, index, device=None, dtype=None, ops: Any = None):
         super().__init__()
         self.self_attn = Attention(config, device=device, dtype=dtype, ops=ops)
         self.mlp = MLP(config, device=device, dtype=dtype, ops=ops)
@@ -226,7 +290,7 @@ class TransformerBlock(nn.Module):
         return x
 
 class TransformerBlockGemma2(nn.Module):
-    def __init__(self, config: Llama2Config, device=None, dtype=None, ops: Any = None):
+    def __init__(self, config: Llama2Config, index, device=None, dtype=None, ops: Any = None):
         super().__init__()
         self.self_attn = Attention(config, device=device, dtype=dtype, ops=ops)
         self.mlp = MLP(config, device=device, dtype=dtype, ops=ops)
@@ -235,6 +299,13 @@ class TransformerBlockGemma2(nn.Module):
         self.pre_feedforward_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
         self.post_feedforward_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
 
+        if config.sliding_attention is not None:  # TODO: implement. (Not that necessary since models are trained on less than 1024 tokens)
+            self.sliding_attention = config.sliding_attention[index % len(config.sliding_attention)]
+        else:
+            self.sliding_attention = False
+
+        self.transformer_type = config.transformer_type
+
     def forward(
         self,
         x: torch.Tensor,
@@ -242,6 +313,14 @@ class TransformerBlockGemma2(nn.Module):
         freqs_cis: Optional[torch.Tensor] = None,
         optimized_attention=None,
     ):
+        if self.transformer_type == 'gemma3':
+            if self.sliding_attention:
+                if x.shape[1] > self.sliding_attention:
+                    logging.warning("Warning: sliding attention not implemented, results may be incorrect")
+                freqs_cis = freqs_cis[1]
+            else:
+                freqs_cis = freqs_cis[0]
+
         # Self Attention
         residual = x
         x = self.input_layernorm(x)
@@ -276,7 +355,7 @@ class Llama2_(nn.Module):
             device=device,
             dtype=dtype
         )
-        if self.config.transformer_type == "gemma2":
+        if self.config.transformer_type == "gemma2" or self.config.transformer_type == "gemma3":
             transformer = TransformerBlockGemma2
             self.normalize_in = True
         else:
@@ -284,8 +363,8 @@ class Llama2_(nn.Module):
             self.normalize_in = False
 
         self.layers = nn.ModuleList([
-            transformer(config, device=device, dtype=dtype, ops=ops)
-            for _ in range(config.num_hidden_layers)
+            transformer(config, index=i, device=device, dtype=dtype, ops=ops)
+            for i in range(config.num_hidden_layers)
         ])
         self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
         # self.lm_head = ops.Linear(config.hidden_size, config.vocab_size, bias=False, device=device, dtype=dtype)
@@ -305,6 +384,7 @@ class Llama2_(nn.Module):
         freqs_cis = precompute_freqs_cis(self.config.head_dim,
                                          position_ids,
                                          self.config.rope_theta,
+                                         self.config.rope_scale,
                                          self.config.rope_dims,
                                          device=x.device)
 
@@ -433,3 +513,12 @@ class Gemma2_2B(BaseLlama, torch.nn.Module):
 
         self.model = Llama2_(config, device=device, dtype=dtype, ops=operations)
         self.dtype = dtype
+
+class Gemma3_4B(BaseLlama, torch.nn.Module):
+    def __init__(self, config_dict, dtype, device, operations):
+        super().__init__()
+        config = Gemma3_4B_Config(**config_dict)
+        self.num_layers = config.num_hidden_layers
+
+        self.model = Llama2_(config, device=device, dtype=dtype, ops=operations)
+        self.dtype = dtype

comfy/text_encoders/lumina2.py

@@ -11,23 +11,41 @@ class Gemma2BTokenizer(sd1_clip.SDTokenizer):
     def state_dict(self):
         return {"spiece_model": self.tokenizer.serialize_model()}
 
+class Gemma3_4BTokenizer(sd1_clip.SDTokenizer):
+    def __init__(self, embedding_directory=None, tokenizer_data={}):
+        tokenizer = tokenizer_data.get("spiece_model", None)
+        super().__init__(tokenizer, pad_with_end=False, embedding_size=2560, embedding_key='gemma3_4b', tokenizer_class=SPieceTokenizer, has_end_token=False, pad_to_max_length=False, max_length=99999999, min_length=1, tokenizer_args={"add_bos": True, "add_eos": False}, tokenizer_data=tokenizer_data)
+
+    def state_dict(self):
+        return {"spiece_model": self.tokenizer.serialize_model()}
 
 class LuminaTokenizer(sd1_clip.SD1Tokenizer):
     def __init__(self, embedding_directory=None, tokenizer_data={}):
         super().__init__(embedding_directory=embedding_directory, tokenizer_data=tokenizer_data, name="gemma2_2b", tokenizer=Gemma2BTokenizer)
 
+class NTokenizer(sd1_clip.SD1Tokenizer):
+    def __init__(self, embedding_directory=None, tokenizer_data={}):
+        super().__init__(embedding_directory=embedding_directory, tokenizer_data=tokenizer_data, name="gemma3_4b", tokenizer=Gemma3_4BTokenizer)
 
 class Gemma2_2BModel(sd1_clip.SDClipModel):
     def __init__(self, device="cpu", layer="hidden", layer_idx=-2, dtype=None, attention_mask=True, model_options={}):
         super().__init__(device=device, layer=layer, layer_idx=layer_idx, textmodel_json_config={}, dtype=dtype, special_tokens={"start": 2, "pad": 0}, layer_norm_hidden_state=False, model_class=comfy.text_encoders.llama.Gemma2_2B, enable_attention_masks=attention_mask, return_attention_masks=attention_mask, model_options=model_options)
 
+class Gemma3_4BModel(sd1_clip.SDClipModel):
+    def __init__(self, device="cpu", layer="hidden", layer_idx=-2, dtype=None, attention_mask=True, model_options={}):
+        super().__init__(device=device, layer=layer, layer_idx=layer_idx, textmodel_json_config={}, dtype=dtype, special_tokens={"start": 2, "pad": 0}, layer_norm_hidden_state=False, model_class=comfy.text_encoders.llama.Gemma3_4B, enable_attention_masks=attention_mask, return_attention_masks=attention_mask, model_options=model_options)
 
 class LuminaModel(sd1_clip.SD1ClipModel):
-    def __init__(self, device="cpu", dtype=None, model_options={}):
-        super().__init__(device=device, dtype=dtype, name="gemma2_2b", clip_model=Gemma2_2BModel, model_options=model_options)
+    def __init__(self, device="cpu", dtype=None, model_options={}, name="gemma2_2b", clip_model=Gemma2_2BModel):
+        super().__init__(device=device, dtype=dtype, name=name, clip_model=clip_model, model_options=model_options)
 
 
-def te(dtype_llama=None, llama_scaled_fp8=None):
+def te(dtype_llama=None, llama_scaled_fp8=None, model_type="gemma2_2b"):
+    if model_type == "gemma2_2b":
+        model = Gemma2_2BModel
+    elif model_type == "gemma3_4b":
+        model = Gemma3_4BModel
+
     class LuminaTEModel_(LuminaModel):
         def __init__(self, device="cpu", dtype=None, model_options={}):
             if llama_scaled_fp8 is not None and "scaled_fp8" not in model_options:
@@ -35,5 +53,5 @@ def te(dtype_llama=None, llama_scaled_fp8=None):
                 model_options["scaled_fp8"] = llama_scaled_fp8
             if dtype_llama is not None:
                 dtype = dtype_llama
-            super().__init__(device=device, dtype=dtype, model_options=model_options)
+            super().__init__(device=device, dtype=dtype, name=model_type, model_options=model_options, clip_model=model)
     return LuminaTEModel_

comfy_api/latest/__init__.py

@@ -8,8 +8,8 @@ from comfy_api.internal.async_to_sync import create_sync_class
 from comfy_api.latest._input import ImageInput, AudioInput, MaskInput, LatentInput, VideoInput
 from comfy_api.latest._input_impl import VideoFromFile, VideoFromComponents
 from comfy_api.latest._util import VideoCodec, VideoContainer, VideoComponents
-from comfy_api.latest._io import _IO as io  #noqa: F401
-from comfy_api.latest._ui import _UI as ui  #noqa: F401
+from . import _io as io
+from . import _ui as ui
 # from comfy_api.latest._resources import _RESOURCES as resources  #noqa: F401
 from comfy_execution.utils import get_executing_context
 from comfy_execution.progress import get_progress_state, PreviewImageTuple
@@ -114,6 +114,8 @@ if TYPE_CHECKING:
     ComfyAPISync: Type[comfy_api.latest.generated.ComfyAPISyncStub.ComfyAPISyncStub]
 ComfyAPISync = create_sync_class(ComfyAPI_latest)
 
+comfy_io = io  # create the new alias for io
+
 __all__ = [
     "ComfyAPI",
     "ComfyAPISync",
@@ -121,4 +123,7 @@ __all__ = [
     "InputImpl",
     "Types",
     "ComfyExtension",
+    "io",
+    "comfy_io",
+    "ui",
 ]

comfy_api/latest/_input/video_types.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 from abc import ABC, abstractmethod
-from typing import Optional, Union
+from typing import Optional, Union, IO
 import io
 import av
 from comfy_api.util import VideoContainer, VideoCodec, VideoComponents
@@ -23,7 +23,7 @@ class VideoInput(ABC):
     @abstractmethod
     def save_to(
         self,
-        path: str,
+        path: Union[str, IO[bytes]],
         format: VideoContainer = VideoContainer.AUTO,
         codec: VideoCodec = VideoCodec.AUTO,
         metadata: Optional[dict] = None

comfy_api/latest/_io.py

@@ -336,11 +336,25 @@ class Combo(ComfyTypeIO):
     class Input(WidgetInput):
         """Combo input (dropdown)."""
         Type = str
-        def __init__(self, id: str, options: list[str]=None, display_name: str=None, optional=False, tooltip: str=None, lazy: bool=None,
-                    default: str=None, control_after_generate: bool=None,
-                    upload: UploadType=None, image_folder: FolderType=None,
-                    remote: RemoteOptions=None,
-                    socketless: bool=None):
+        def __init__(
+            self,
+            id: str,
+            options: list[str] | list[int] | type[Enum] = None,
+            display_name: str=None,
+            optional=False,
+            tooltip: str=None,
+            lazy: bool=None,
+            default: str | int | Enum = None,
+            control_after_generate: bool=None,
+            upload: UploadType=None,
+            image_folder: FolderType=None,
+            remote: RemoteOptions=None,
+            socketless: bool=None,
+        ):
+            if isinstance(options, type) and issubclass(options, Enum):
+                options = [v.value for v in options]
+            if isinstance(default, Enum):
+                default = default.value
             super().__init__(id, display_name, optional, tooltip, lazy, default, socketless)
             self.multiselect = False
             self.options = options
@@ -1568,79 +1582,78 @@ class _UIOutput(ABC):
         ...
 
 
-class _IO:
-    FolderType = FolderType
-    UploadType = UploadType
-    RemoteOptions = RemoteOptions
-    NumberDisplay = NumberDisplay
+__all__ = [
+    "FolderType",
+    "UploadType",
+    "RemoteOptions",
+    "NumberDisplay",
 
-    comfytype = staticmethod(comfytype)
-    Custom = staticmethod(Custom)
-    Input = Input
-    WidgetInput = WidgetInput
-    Output = Output
-    ComfyTypeI = ComfyTypeI
-    ComfyTypeIO = ComfyTypeIO
-    #---------------------------------
+    "comfytype",
+    "Custom",
+    "Input",
+    "WidgetInput",
+    "Output",
+    "ComfyTypeI",
+    "ComfyTypeIO",
     # Supported Types
-    Boolean = Boolean
-    Int = Int
-    Float = Float
-    String = String
-    Combo = Combo
-    MultiCombo = MultiCombo
-    Image = Image
-    WanCameraEmbedding = WanCameraEmbedding
-    Webcam = Webcam
-    Mask = Mask
-    Latent = Latent
-    Conditioning = Conditioning
-    Sampler = Sampler
-    Sigmas = Sigmas
-    Noise = Noise
-    Guider = Guider
-    Clip = Clip
-    ControlNet = ControlNet
-    Vae = Vae
-    Model = Model
-    ModelPatch = ModelPatch
-    ClipVision = ClipVision
-    ClipVisionOutput = ClipVisionOutput
-    AudioEncoder = AudioEncoder
-    AudioEncoderOutput = AudioEncoderOutput
-    StyleModel = StyleModel
-    Gligen = Gligen
-    UpscaleModel = UpscaleModel
-    Audio = Audio
-    Video = Video
-    SVG = SVG
-    LoraModel = LoraModel
-    LossMap = LossMap
-    Voxel = Voxel
-    Mesh = Mesh
-    Hooks = Hooks
-    HookKeyframes = HookKeyframes
-    TimestepsRange = TimestepsRange
-    LatentOperation = LatentOperation
-    FlowControl = FlowControl
-    Accumulation = Accumulation
-    Load3DCamera = Load3DCamera
-    Load3D = Load3D
-    Load3DAnimation = Load3DAnimation
-    Photomaker = Photomaker
-    Point = Point
-    FaceAnalysis = FaceAnalysis
-    BBOX = BBOX
-    SEGS = SEGS
-    AnyType = AnyType
-    MultiType = MultiType
-    #---------------------------------
-    HiddenHolder = HiddenHolder
-    Hidden = Hidden
-    NodeInfoV1 = NodeInfoV1
-    NodeInfoV3 = NodeInfoV3
-    Schema = Schema
-    ComfyNode = ComfyNode
-    NodeOutput = NodeOutput
-    add_to_dict_v1 = staticmethod(add_to_dict_v1)
-    add_to_dict_v3 = staticmethod(add_to_dict_v3)
+    "Boolean",
+    "Int",
+    "Float",
+    "String",
+    "Combo",
+    "MultiCombo",
+    "Image",
+    "WanCameraEmbedding",
+    "Webcam",
+    "Mask",
+    "Latent",
+    "Conditioning",
+    "Sampler",
+    "Sigmas",
+    "Noise",
+    "Guider",
+    "Clip",
+    "ControlNet",
+    "Vae",
+    "Model",
+    "ClipVision",
+    "ClipVisionOutput",
+    "AudioEncoder",
+    "AudioEncoderOutput",
+    "StyleModel",
+    "Gligen",
+    "UpscaleModel",
+    "Audio",
+    "Video",
+    "SVG",
+    "LoraModel",
+    "LossMap",
+    "Voxel",
+    "Mesh",
+    "Hooks",
+    "HookKeyframes",
+    "TimestepsRange",
+    "LatentOperation",
+    "FlowControl",
+    "Accumulation",
+    "Load3DCamera",
+    "Load3D",
+    "Load3DAnimation",
+    "Photomaker",
+    "Point",
+    "FaceAnalysis",
+    "BBOX",
+    "SEGS",
+    "AnyType",
+    "MultiType",
+    # Other classes
+    "HiddenHolder",
+    "Hidden",
+    "NodeInfoV1",
+    "NodeInfoV3",
+    "Schema",
+    "ComfyNode",
+    "NodeOutput",
+    "add_to_dict_v1",
+    "add_to_dict_v3",
+]

comfy_api/latest/_ui.py

@@ -449,15 +449,16 @@ class PreviewText(_UIOutput):
         return {"text": (self.value,)}
 
 
-class _UI:
-    SavedResult = SavedResult
-    SavedImages = SavedImages
-    SavedAudios = SavedAudios
-    ImageSaveHelper = ImageSaveHelper
-    AudioSaveHelper = AudioSaveHelper
-    PreviewImage = PreviewImage
-    PreviewMask = PreviewMask
-    PreviewAudio = PreviewAudio
-    PreviewVideo = PreviewVideo
-    PreviewUI3D = PreviewUI3D
-    PreviewText = PreviewText
+__all__ = [
+    "SavedResult",
+    "SavedImages",
+    "SavedAudios",
+    "ImageSaveHelper",
+    "AudioSaveHelper",
+    "PreviewImage",
+    "PreviewMask",
+    "PreviewAudio",
+    "PreviewVideo",
+    "PreviewUI3D",
+    "PreviewText",
+]

comfy_api_nodes/apinode_utils.py

@@ -18,7 +18,7 @@ from comfy_api_nodes.apis.client import (
     UploadResponse,
 )
 from server import PromptServer
-
+from comfy.cli_args import args
 
 import numpy as np
 from PIL import Image
@@ -30,7 +30,9 @@ from io import BytesIO
 import av
 
 
-async def download_url_to_video_output(video_url: str, timeout: int = None) -> VideoFromFile:
+async def download_url_to_video_output(
+    video_url: str, timeout: int = None, auth_kwargs: Optional[dict[str, str]] = None
+) -> VideoFromFile:
     """Downloads a video from a URL and returns a `VIDEO` output.
 
     Args:
@@ -39,7 +41,7 @@ async def download_url_to_video_output(video_url: str, timeout: int = None) -> V
     Returns:
         A Comfy node `VIDEO` output.
     """
-    video_io = await download_url_to_bytesio(video_url, timeout)
+    video_io = await download_url_to_bytesio(video_url, timeout, auth_kwargs=auth_kwargs)
     if video_io is None:
         error_msg = f"Failed to download video from {video_url}"
         logging.error(error_msg)
@@ -152,7 +154,7 @@ def validate_aspect_ratio(
             raise TypeError(
                 f"Aspect ratio cannot reduce to any less than {minimum_ratio_str} ({minimum_ratio}), but was {aspect_ratio} ({calculated_ratio})."
             )
-        elif calculated_ratio > maximum_ratio:
+        if calculated_ratio > maximum_ratio:
             raise TypeError(
                 f"Aspect ratio cannot reduce to any greater than {maximum_ratio_str} ({maximum_ratio}), but was {aspect_ratio} ({calculated_ratio})."
             )
@@ -164,7 +166,9 @@ def mimetype_to_extension(mime_type: str) -> str:
     return mime_type.split("/")[-1].lower()
 
 
-async def download_url_to_bytesio(url: str, timeout: int = None) -> BytesIO:
+async def download_url_to_bytesio(
+    url: str, timeout: int = None, auth_kwargs: Optional[dict[str, str]] = None
+) -> BytesIO:
     """Downloads content from a URL using requests and returns it as BytesIO.
 
     Args:
@@ -174,9 +178,18 @@ async def download_url_to_bytesio(url: str, timeout: int = None) -> BytesIO:
     Returns:
         BytesIO object containing the downloaded content.
     """
+    headers = {}
+    if url.startswith("/proxy/"):
+        url = str(args.comfy_api_base).rstrip("/") + url
+        auth_token = auth_kwargs.get("auth_token")
+        comfy_api_key = auth_kwargs.get("comfy_api_key")
+        if auth_token:
+            headers["Authorization"] = f"Bearer {auth_token}"
+        elif comfy_api_key:
+            headers["X-API-KEY"] = comfy_api_key
     timeout_cfg = aiohttp.ClientTimeout(total=timeout) if timeout else None
     async with aiohttp.ClientSession(timeout=timeout_cfg) as session:
-        async with session.get(url) as resp:
+        async with session.get(url, headers=headers) as resp:
             resp.raise_for_status()  # Raises HTTPError for bad responses (4XX or 5XX)
             return BytesIO(await resp.read())
 
@@ -256,7 +269,7 @@ def tensor_to_bytesio(
         mime_type: Target image MIME type (e.g., 'image/png', 'image/jpeg', 'image/webp', 'video/mp4').
 
     Returns:
-        Named BytesIO object containing the image data.
+        Named BytesIO object containing the image data, with pointer set to the start of buffer.
     """
     if not mime_type:
         mime_type = "image/png"
@@ -418,7 +431,7 @@ async def upload_video_to_comfyapi(
                     f"Video duration ({actual_duration:.2f}s) exceeds the maximum allowed ({max_duration}s)."
                 )
         except Exception as e:
-            logging.error(f"Error getting video duration: {e}")
+            logging.error("Error getting video duration: %s", str(e))
             raise ValueError(f"Could not verify video duration from source: {e}") from e
 
     upload_mime_type = f"video/{container.value.lower()}"

comfy_api_nodes/apis/__init__.py

@@ -1321,6 +1321,7 @@ class KlingTextToVideoModelName(str, Enum):
     kling_v1 = 'kling-v1'
     kling_v1_6 = 'kling-v1-6'
     kling_v2_1_master = 'kling-v2-1-master'
+    kling_v2_5_turbo = 'kling-v2-5-turbo'
 
 
 class KlingVideoGenAspectRatio(str, Enum):
@@ -1355,6 +1356,7 @@ class KlingVideoGenModelName(str, Enum):
     kling_v2_master = 'kling-v2-master'
     kling_v2_1 = 'kling-v2-1'
     kling_v2_1_master = 'kling-v2-1-master'
+    kling_v2_5_turbo = 'kling-v2-5-turbo'
 
 
 class KlingVideoResult(BaseModel):

comfy_api_nodes/apis/client.py

@@ -98,7 +98,7 @@ import io
 import os
 import socket
 from aiohttp.client_exceptions import ClientError, ClientResponseError
-from typing import Dict, Type, Optional, Any, TypeVar, Generic, Callable, Tuple
+from typing import Type, Optional, Any, TypeVar, Generic, Callable
 from enum import Enum
 import json
 from urllib.parse import urljoin, urlparse
@@ -175,7 +175,7 @@ class ApiClient:
         max_retries: int = 3,
         retry_delay: float = 1.0,
         retry_backoff_factor: float = 2.0,
-        retry_status_codes: Optional[Tuple[int, ...]] = None,
+        retry_status_codes: Optional[tuple[int, ...]] = None,
         session: Optional[aiohttp.ClientSession] = None,
     ):
         self.base_url = base_url
@@ -199,9 +199,9 @@ class ApiClient:
 
     @staticmethod
     def _create_json_payload_args(
-        data: Optional[Dict[str, Any]] = None,
-        headers: Optional[Dict[str, str]] = None,
-    ) -> Dict[str, Any]:
+        data: Optional[dict[str, Any]] = None,
+        headers: Optional[dict[str, str]] = None,
+    ) -> dict[str, Any]:
         return {
             "json": data,
             "headers": headers,
@@ -209,24 +209,27 @@ class ApiClient:
 
     def _create_form_data_args(
         self,
-        data: Dict[str, Any] | None,
-        files: Dict[str, Any] | None,
-        headers: Optional[Dict[str, str]] = None,
+        data: dict[str, Any] | None,
+        files: dict[str, Any] | None,
+        headers: Optional[dict[str, str]] = None,
         multipart_parser: Callable | None = None,
-    ) -> Dict[str, Any]:
+    ) -> dict[str, Any]:
         if headers and "Content-Type" in headers:
             del headers["Content-Type"]
 
         if multipart_parser and data:
             data = multipart_parser(data)
 
-        form = aiohttp.FormData(default_to_multipart=True)
-        if data:  # regular text fields
-            for k, v in data.items():
-                if v is None:
-                    continue  # aiohttp fails to serialize "None" values
-                # aiohttp expects strings or bytes; convert enums etc.
-                form.add_field(k, str(v) if not isinstance(v, (bytes, bytearray)) else v)
+        if isinstance(data, aiohttp.FormData):
+            form = data  # If the parser already returned a FormData, pass it through
+        else:
+            form = aiohttp.FormData(default_to_multipart=True)
+            if data:  # regular text fields
+                for k, v in data.items():
+                    if v is None:
+                        continue  # aiohttp fails to serialize "None" values
+                    # aiohttp expects strings or bytes; convert enums etc.
+                    form.add_field(k, str(v) if not isinstance(v, (bytes, bytearray)) else v)
 
         if files:
             file_iter = files if isinstance(files, list) else files.items()
@@ -251,9 +254,9 @@ class ApiClient:
 
     @staticmethod
     def _create_urlencoded_form_data_args(
-        data: Dict[str, Any],
-        headers: Optional[Dict[str, str]] = None,
-    ) -> Dict[str, Any]:
+        data: dict[str, Any],
+        headers: Optional[dict[str, str]] = None,
+    ) -> dict[str, Any]:
         headers = headers or {}
         headers["Content-Type"] = "application/x-www-form-urlencoded"
         return {
@@ -261,7 +264,7 @@ class ApiClient:
             "headers": headers,
         }
 
-    def get_headers(self) -> Dict[str, str]:
+    def get_headers(self) -> dict[str, str]:
         """Get headers for API requests, including authentication if available"""
         headers = {"Content-Type": "application/json", "Accept": "application/json"}
 
@@ -272,7 +275,7 @@ class ApiClient:
 
         return headers
 
-    async def _check_connectivity(self, target_url: str) -> Dict[str, bool]:
+    async def _check_connectivity(self, target_url: str) -> dict[str, bool]:
         """
         Check connectivity to determine if network issues are local or server-related.
 
@@ -313,14 +316,14 @@ class ApiClient:
         self,
         method: str,
         path: str,
-        params: Optional[Dict[str, Any]] = None,
-        data: Optional[Dict[str, Any]] = None,
-        files: Optional[Dict[str, Any] | list[tuple[str, Any]]] = None,
-        headers: Optional[Dict[str, str]] = None,
+        params: Optional[dict[str, Any]] = None,
+        data: Optional[dict[str, Any]] = None,
+        files: Optional[dict[str, Any] | list[tuple[str, Any]]] = None,
+        headers: Optional[dict[str, str]] = None,
         content_type: str = "application/json",
         multipart_parser: Callable | None = None,
         retry_count: int = 0,  # Used internally for tracking retries
-    ) -> Dict[str, Any]:
+    ) -> dict[str, Any]:
         """
         Make an HTTP request to the API with automatic retries for transient errors.
 
@@ -356,10 +359,10 @@ class ApiClient:
         if params:
             params = {k: v for k, v in params.items() if v is not None}  # aiohttp fails to serialize None values
 
-        logging.debug(f"[DEBUG] Request Headers: {request_headers}")
-        logging.debug(f"[DEBUG] Files: {files}")
-        logging.debug(f"[DEBUG] Params: {params}")
-        logging.debug(f"[DEBUG] Data: {data}")
+        logging.debug("[DEBUG] Request Headers: %s", request_headers)
+        logging.debug("[DEBUG] Files: %s", files)
+        logging.debug("[DEBUG] Params: %s", params)
+        logging.debug("[DEBUG] Data: %s", data)
 
         if content_type == "application/x-www-form-urlencoded":
             payload_args = self._create_urlencoded_form_data_args(data or {}, request_headers)
@@ -482,7 +485,7 @@ class ApiClient:
             retry_delay: Initial delay between retries in seconds
             retry_backoff_factor: Multiplier for the delay after each retry
         """
-        headers: Dict[str, str] = {}
+        headers: dict[str, str] = {}
         skip_auto_headers: set[str] = set()
         if content_type:
             headers["Content-Type"] = content_type
@@ -555,7 +558,7 @@ class ApiClient:
         *req_meta,
         retry_count: int,
         response_content: dict | str = "",
-    ) -> Dict[str, Any]:
+    ) -> dict[str, Any]:
         status_code = exc.status
         if status_code == 401:
             user_friendly = "Unauthorized: Please login first to use this node."
@@ -589,9 +592,9 @@ class ApiClient:
             error_message=f"HTTP Error {exc.status}",
         )
 
-        logging.debug(f"[DEBUG] API Error: {user_friendly} (Status: {status_code})")
+        logging.debug("[DEBUG] API Error: %s (Status: %s)", user_friendly, status_code)
         if response_content:
-            logging.debug(f"[DEBUG] Response content: {response_content}")
+            logging.debug("[DEBUG] Response content: %s", response_content)
 
         # Retry if eligible
         if status_code in self.retry_status_codes and retry_count < self.max_retries:
@@ -656,7 +659,7 @@ class ApiEndpoint(Generic[T, R]):
         method: HttpMethod,
         request_model: Type[T],
         response_model: Type[R],
-        query_params: Optional[Dict[str, Any]] = None,
+        query_params: Optional[dict[str, Any]] = None,
     ):
         """Initialize an API endpoint definition.
 
@@ -681,11 +684,11 @@ class SynchronousOperation(Generic[T, R]):
         self,
         endpoint: ApiEndpoint[T, R],
         request: T,
-        files: Optional[Dict[str, Any] | list[tuple[str, Any]]] = None,
+        files: Optional[dict[str, Any] | list[tuple[str, Any]]] = None,
         api_base: str | None = None,
         auth_token: Optional[str] = None,
         comfy_api_key: Optional[str] = None,
-        auth_kwargs: Optional[Dict[str, str]] = None,
+        auth_kwargs: Optional[dict[str, str]] = None,
         timeout: float = 7200.0,
         verify_ssl: bool = True,
         content_type: str = "application/json",
@@ -726,7 +729,7 @@ class SynchronousOperation(Generic[T, R]):
             )
 
         try:
-            request_dict: Optional[Dict[str, Any]]
+            request_dict: Optional[dict[str, Any]]
             if isinstance(self.request, EmptyRequest):
                 request_dict = None
             else:
@@ -735,11 +738,9 @@ class SynchronousOperation(Generic[T, R]):
                     if isinstance(v, Enum):
                         request_dict[k] = v.value
 
-            logging.debug(
-                f"[DEBUG] API Request: {self.endpoint.method.value} {self.endpoint.path}"
-            )
-            logging.debug(f"[DEBUG] Request Data: {json.dumps(request_dict, indent=2)}")
-            logging.debug(f"[DEBUG] Query Params: {self.endpoint.query_params}")
+            logging.debug("[DEBUG] API Request: %s %s", self.endpoint.method.value, self.endpoint.path)
+            logging.debug("[DEBUG] Request Data: %s", json.dumps(request_dict, indent=2))
+            logging.debug("[DEBUG] Query Params: %s", self.endpoint.query_params)
 
             response_json = await client.request(
                 self.endpoint.method.value,
@@ -754,11 +755,11 @@ class SynchronousOperation(Generic[T, R]):
             logging.debug("=" * 50)
             logging.debug("[DEBUG] RESPONSE DETAILS:")
             logging.debug("[DEBUG] Status Code: 200 (Success)")
-            logging.debug(f"[DEBUG] Response Body: {json.dumps(response_json, indent=2)}")
+            logging.debug("[DEBUG] Response Body: %s", json.dumps(response_json, indent=2))
             logging.debug("=" * 50)
 
             parsed_response = self.endpoint.response_model.model_validate(response_json)
-            logging.debug(f"[DEBUG] Parsed Response: {parsed_response}")
+            logging.debug("[DEBUG] Parsed Response: %s", parsed_response)
             return parsed_response
         finally:
             if owns_client:
@@ -781,14 +782,14 @@ class PollingOperation(Generic[T, R]):
         poll_endpoint: ApiEndpoint[EmptyRequest, R],
         completed_statuses: list[str],
         failed_statuses: list[str],
-        status_extractor: Callable[[R], str],
-        progress_extractor: Callable[[R], float] | None = None,
-        result_url_extractor: Callable[[R], str] | None = None,
+        status_extractor: Callable[[R], Optional[str]],
+        progress_extractor: Callable[[R], Optional[float]] | None = None,
+        result_url_extractor: Callable[[R], Optional[str]] | None = None,
         request: Optional[T] = None,
         api_base: str | None = None,
         auth_token: Optional[str] = None,
         comfy_api_key: Optional[str] = None,
-        auth_kwargs: Optional[Dict[str, str]] = None,
+        auth_kwargs: Optional[dict[str, str]] = None,
         poll_interval: float = 5.0,
         max_poll_attempts: int = 120,  # Default max polling attempts (10 minutes with 5s interval)
         max_retries: int = 3,  # Max retries per individual API call
@@ -874,7 +875,7 @@ class PollingOperation(Generic[T, R]):
         status = TaskStatus.PENDING
         for poll_count in range(1, self.max_poll_attempts + 1):
             try:
-                logging.debug(f"[DEBUG] Polling attempt #{poll_count}")
+                logging.debug("[DEBUG] Polling attempt #%s", poll_count)
 
                 request_dict = (
                     None if self.request is None else self.request.model_dump(exclude_none=True)
@@ -882,10 +883,13 @@ class PollingOperation(Generic[T, R]):
 
                 if poll_count == 1:
                     logging.debug(
-                        f"[DEBUG] Poll Request: {self.poll_endpoint.method.value} {self.poll_endpoint.path}"
+                        "[DEBUG] Poll Request: %s %s",
+                        self.poll_endpoint.method.value,
+                        self.poll_endpoint.path,
                     )
                     logging.debug(
-                        f"[DEBUG] Poll Request Data: {json.dumps(request_dict, indent=2) if request_dict else 'None'}"
+                        "[DEBUG] Poll Request Data: %s",
+                        json.dumps(request_dict, indent=2) if request_dict else "None",
                     )
 
                 # Query task status
@@ -900,7 +904,7 @@ class PollingOperation(Generic[T, R]):
 
                 # Check if task is complete
                 status = self._check_task_status(response_obj)
-                logging.debug(f"[DEBUG] Task Status: {status}")
+                logging.debug("[DEBUG] Task Status: %s", status)
 
                 # If progress extractor is provided, extract progress
                 if self.progress_extractor:
@@ -914,7 +918,7 @@ class PollingOperation(Generic[T, R]):
                         result_url = self.result_url_extractor(response_obj)
                         if result_url:
                             message = f"Result URL: {result_url}"
-                    logging.debug(f"[DEBUG] {message}")
+                    logging.debug("[DEBUG] %s", message)
                     self._display_text_on_node(message)
                     self.final_response = response_obj
                     if self.progress_extractor:
@@ -922,7 +926,7 @@ class PollingOperation(Generic[T, R]):
                     return self.final_response
                 if status == TaskStatus.FAILED:
                     message = f"Task failed: {json.dumps(resp)}"
-                    logging.error(f"[DEBUG] {message}")
+                    logging.error("[DEBUG] %s", message)
                     raise Exception(message)
                 logging.debug("[DEBUG] Task still pending, continuing to poll...")
                 # Task pending – wait
@@ -936,7 +940,12 @@ class PollingOperation(Generic[T, R]):
                     raise Exception(
                         f"Polling aborted after {consecutive_errors} network errors: {str(e)}"
                     ) from e
-                logging.warning("Network error (%s/%s): %s", consecutive_errors, max_consecutive_errors, str(e))
+                logging.warning(
+                    "Network error (%s/%s): %s",
+                    consecutive_errors,
+                    max_consecutive_errors,
+                    str(e),
+                )
                 await asyncio.sleep(self.poll_interval)
             except Exception as e:
                 # For other errors, increment count and potentially abort
@@ -946,10 +955,13 @@ class PollingOperation(Generic[T, R]):
                         f"Polling aborted after {consecutive_errors} consecutive errors: {str(e)}"
                     ) from e
 
-                logging.error(f"[DEBUG] Polling error: {str(e)}")
+                logging.error("[DEBUG] Polling error: %s", str(e))
                 logging.warning(
-                    f"Error during polling (attempt {poll_count}/{self.max_poll_attempts}): {str(e)}. "
-                    f"Will retry in {self.poll_interval} seconds."
+                    "Error during polling (attempt %s/%s): %s. Will retry in %s seconds.",
+                    poll_count,
+                    self.max_poll_attempts,
+                    str(e),
+                    self.poll_interval,
                 )
                 await asyncio.sleep(self.poll_interval)
 

comfy_api_nodes/apis/pika_defs.py

@@ -0,0 +1,100 @@
+from typing import Optional
+from enum import Enum
+from pydantic import BaseModel, Field
+
+
+class Pikaffect(str, Enum):
+    Cake_ify = "Cake-ify"
+    Crumble = "Crumble"
+    Crush = "Crush"
+    Decapitate = "Decapitate"
+    Deflate = "Deflate"
+    Dissolve = "Dissolve"
+    Explode = "Explode"
+    Eye_pop = "Eye-pop"
+    Inflate = "Inflate"
+    Levitate = "Levitate"
+    Melt = "Melt"
+    Peel = "Peel"
+    Poke = "Poke"
+    Squish = "Squish"
+    Ta_da = "Ta-da"
+    Tear = "Tear"
+
+
+class PikaBodyGenerate22C2vGenerate22PikascenesPost(BaseModel):
+    aspectRatio: Optional[float] = Field(None, description='Aspect ratio (width / height)')
+    duration: Optional[int] = Field(5)
+    ingredientsMode: str = Field(...)
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    resolution: Optional[str] = Field('1080p')
+    seed: Optional[int] = Field(None)
+
+
+class PikaGenerateResponse(BaseModel):
+    video_id: str = Field(...)
+
+
+class PikaBodyGenerate22I2vGenerate22I2vPost(BaseModel):
+    duration: Optional[int] = 5
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    resolution: Optional[str] = '1080p'
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGenerate22KeyframeGenerate22PikaframesPost(BaseModel):
+    duration: Optional[int] = Field(None, ge=5, le=10)
+    negativePrompt: Optional[str] = Field(None)
+    promptText: str = Field(...)
+    resolution: Optional[str] = '1080p'
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGenerate22T2vGenerate22T2vPost(BaseModel):
+    aspectRatio: Optional[float] = Field(
+        1.7777777777777777,
+        description='Aspect ratio (width / height)',
+        ge=0.4,
+        le=2.5,
+    )
+    duration: Optional[int] = 5
+    negativePrompt: Optional[str] = Field(None)
+    promptText: str = Field(...)
+    resolution: Optional[str] = '1080p'
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGeneratePikadditionsGeneratePikadditionsPost(BaseModel):
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGeneratePikaffectsGeneratePikaffectsPost(BaseModel):
+    negativePrompt: Optional[str] = Field(None)
+    pikaffect: Optional[str] = None
+    promptText: Optional[str] = Field(None)
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGeneratePikaswapsGeneratePikaswapsPost(BaseModel):
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    seed: Optional[int] = Field(None)
+    modifyRegionRoi: Optional[str] = Field(None)
+
+
+class PikaStatusEnum(str, Enum):
+    queued = "queued"
+    started = "started"
+    finished = "finished"
+    failed = "failed"
+
+
+class PikaVideoResponse(BaseModel):
+    id: str = Field(...)
+    progress: Optional[int] = Field(None)
+    status: PikaStatusEnum
+    url: Optional[str] = Field(None)

comfy_api_nodes/apis/request_logger.py

@@ -21,7 +21,7 @@ def get_log_directory():
     try:
         os.makedirs(log_dir, exist_ok=True)
     except Exception as e:
-        logger.error(f"Error creating API log directory {log_dir}: {e}")
+        logger.error("Error creating API log directory %s: %s", log_dir, str(e))
         # Fallback to base temp directory if sub-directory creation fails
         return base_temp_dir
     return log_dir
@@ -122,9 +122,9 @@ def log_request_response(
     try:
         with open(filepath, "w", encoding="utf-8") as f:
             f.write("\n".join(log_content))
-        logger.debug(f"API log saved to: {filepath}")
+        logger.debug("API log saved to: %s", filepath)
     except Exception as e:
-        logger.error(f"Error writing API log to {filepath}: {e}")
+        logger.error("Error writing API log to %s: %s", filepath, str(e))
 
 
 if __name__ == '__main__':

comfy_api_nodes/nodes_bytedance.py

@@ -249,8 +249,8 @@ class ByteDanceImageNode(comfy_io.ComfyNode):
             inputs=[
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in Text2ImageModelName],
-                    default=Text2ImageModelName.seedream_3.value,
+                    options=Text2ImageModelName,
+                    default=Text2ImageModelName.seedream_3,
                     tooltip="Model name",
                 ),
                 comfy_io.String.Input(
@@ -382,8 +382,8 @@ class ByteDanceImageEditNode(comfy_io.ComfyNode):
             inputs=[
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in Image2ImageModelName],
-                    default=Image2ImageModelName.seededit_3.value,
+                    options=Image2ImageModelName,
+                    default=Image2ImageModelName.seededit_3,
                     tooltip="Model name",
                 ),
                 comfy_io.Image.Input(
@@ -676,8 +676,8 @@ class ByteDanceTextToVideoNode(comfy_io.ComfyNode):
             inputs=[
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in Text2VideoModelName],
-                    default=Text2VideoModelName.seedance_1_pro.value,
+                    options=Text2VideoModelName,
+                    default=Text2VideoModelName.seedance_1_pro,
                     tooltip="Model name",
                 ),
                 comfy_io.String.Input(
@@ -793,8 +793,8 @@ class ByteDanceImageToVideoNode(comfy_io.ComfyNode):
             inputs=[
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in Image2VideoModelName],
-                    default=Image2VideoModelName.seedance_1_pro.value,
+                    options=Image2VideoModelName,
+                    default=Image2VideoModelName.seedance_1_pro,
                     tooltip="Model name",
                 ),
                 comfy_io.String.Input(

comfy_api_nodes/nodes_gemini.py

@@ -39,6 +39,7 @@ from comfy_api_nodes.apinode_utils import (
     tensor_to_base64_string,
     bytesio_to_image_tensor,
 )
+from comfy_api.util import VideoContainer, VideoCodec
 
 
 GEMINI_BASE_ENDPOINT = "/proxy/vertexai/gemini"
@@ -310,7 +311,7 @@ class GeminiNode(ComfyNodeABC):
         Returns:
             List of GeminiPart objects containing the encoded video.
         """
-        from comfy_api.util import VideoContainer, VideoCodec
+
         base_64_string = video_to_base64_string(
             video_input,
             container_format=VideoContainer.MP4,
@@ -490,7 +491,6 @@ class GeminiInputFiles(ComfyNodeABC):
         # Use base64 string directly, not the data URI
         with open(file_path, "rb") as f:
             file_content = f.read()
-        import base64
         base64_str = base64.b64encode(file_content).decode("utf-8")
 
         return GeminiPart(

comfy_api_nodes/nodes_luma.py

@@ -181,11 +181,11 @@ class LumaImageGenerationNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaImageModel],
+                    options=LumaImageModel,
                 ),
                 comfy_io.Combo.Input(
                     "aspect_ratio",
-                    options=[ratio.value for ratio in LumaAspectRatio],
+                    options=LumaAspectRatio,
                     default=LumaAspectRatio.ratio_16_9,
                 ),
                 comfy_io.Int.Input(
@@ -366,7 +366,7 @@ class LumaImageModifyNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaImageModel],
+                    options=LumaImageModel,
                 ),
                 comfy_io.Int.Input(
                     "seed",
@@ -466,21 +466,21 @@ class LumaTextToVideoGenerationNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaVideoModel],
+                    options=LumaVideoModel,
                 ),
                 comfy_io.Combo.Input(
                     "aspect_ratio",
-                    options=[ratio.value for ratio in LumaAspectRatio],
+                    options=LumaAspectRatio,
                     default=LumaAspectRatio.ratio_16_9,
                 ),
                 comfy_io.Combo.Input(
                     "resolution",
-                    options=[resolution.value for resolution in LumaVideoOutputResolution],
+                    options=LumaVideoOutputResolution,
                     default=LumaVideoOutputResolution.res_540p,
                 ),
                 comfy_io.Combo.Input(
                     "duration",
-                    options=[dur.value for dur in LumaVideoModelOutputDuration],
+                    options=LumaVideoModelOutputDuration,
                 ),
                 comfy_io.Boolean.Input(
                     "loop",
@@ -595,7 +595,7 @@ class LumaImageToVideoGenerationNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaVideoModel],
+                    options=LumaVideoModel,
                 ),
                 # comfy_io.Combo.Input(
                 #     "aspect_ratio",
@@ -604,7 +604,7 @@ class LumaImageToVideoGenerationNode(comfy_io.ComfyNode):
                 # ),
                 comfy_io.Combo.Input(
                     "resolution",
-                    options=[resolution.value for resolution in LumaVideoOutputResolution],
+                    options=LumaVideoOutputResolution,
                     default=LumaVideoOutputResolution.res_540p,
                 ),
                 comfy_io.Combo.Input(

comfy_api_nodes/nodes_minimax.py

@@ -500,7 +500,7 @@ class MinimaxHailuoVideoNode(comfy_io.ComfyNode):
             raise Exception(
                 f"No video was found in the response. Full response: {file_result.model_dump()}"
             )
-        logging.info(f"Generated video URL: {file_url}")
+        logging.info("Generated video URL: %s", file_url)
         if cls.hidden.unique_id:
             if hasattr(file_result.file, "backup_download_url"):
                 message = f"Result URL: {file_url}\nBackup URL: {file_result.file.backup_download_url}"

comfy_api_nodes/nodes_moonvalley.py

@@ -2,11 +2,7 @@ import logging
 from typing import Any, Callable, Optional, TypeVar
 import torch
 from typing_extensions import override
-from comfy_api_nodes.util.validation_utils import (
-    get_image_dimensions,
-    validate_image_dimensions,
-)
-
+from comfy_api_nodes.util.validation_utils import validate_image_dimensions
 
 from comfy_api_nodes.apis import (
     MoonvalleyTextToVideoRequest,
@@ -132,47 +128,6 @@ def validate_prompts(
     return True
 
 
-def validate_input_media(width, height, with_frame_conditioning, num_frames_in=None):
-    # inference validation
-    # T = num_frames
-    # in all cases, the following must be true: T divisible by 16 and H,W by 8. in addition...
-    # with image conditioning: H*W must be divisible by 8192
-    # without image conditioning: T divisible by 32
-    if num_frames_in and not num_frames_in % 16 == 0:
-        return False, ("The input video total frame count must be divisible by 16!")
-
-    if height % 8 != 0 or width % 8 != 0:
-        return False, (
-            f"Height ({height}) and width ({width}) must be " "divisible by 8"
-        )
-
-    if with_frame_conditioning:
-        if (height * width) % 8192 != 0:
-            return False, (
-                f"Height * width ({height * width}) must be "
-                "divisible by 8192 for frame conditioning"
-            )
-    else:
-        if num_frames_in and not num_frames_in % 32 == 0:
-            return False, ("The input video total frame count must be divisible by 32!")
-
-
-def validate_input_image(
-    image: torch.Tensor, with_frame_conditioning: bool = False
-) -> None:
-    """
-    Validates the input image adheres to the expectations of the API:
-    - The image resolution should not be less than 300*300px
-    - The aspect ratio of the image should be between 1:2.5 ~ 2.5:1
-
-    """
-    height, width = get_image_dimensions(image)
-    validate_input_media(width, height, with_frame_conditioning)
-    validate_image_dimensions(
-        image, min_width=300, min_height=300, max_height=MAX_HEIGHT, max_width=MAX_WIDTH
-    )
-
-
 def validate_video_to_video_input(video: VideoInput) -> VideoInput:
     """
     Validates and processes video input for Moonvalley Video-to-Video generation.
@@ -282,7 +237,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
         audio_stream = None
 
         for stream in input_container.streams:
-            logging.info(f"Found stream: type={stream.type}, class={type(stream)}")
+            logging.info("Found stream: type=%s, class=%s", stream.type, type(stream))
             if isinstance(stream, av.VideoStream):
                 # Create output video stream with same parameters
                 video_stream = output_container.add_stream(
@@ -292,7 +247,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
                 video_stream.height = stream.height
                 video_stream.pix_fmt = "yuv420p"
                 logging.info(
-                    f"Added video stream: {stream.width}x{stream.height} @ {stream.average_rate}fps"
+                    "Added video stream: %sx%s @ %sfps", stream.width, stream.height, stream.average_rate
                 )
             elif isinstance(stream, av.AudioStream):
                 # Create output audio stream with same parameters
@@ -301,9 +256,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
                 )
                 audio_stream.sample_rate = stream.sample_rate
                 audio_stream.layout = stream.layout
-                logging.info(
-                    f"Added audio stream: {stream.sample_rate}Hz, {stream.channels} channels"
-                )
+                logging.info("Added audio stream: %sHz, %s channels", stream.sample_rate, stream.channels)
 
         # Calculate target frame count that's divisible by 16
         fps = input_container.streams.video[0].average_rate
@@ -333,9 +286,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
             for packet in video_stream.encode():
                 output_container.mux(packet)
 
-            logging.info(
-                f"Encoded {frame_count} video frames (target: {target_frames})"
-            )
+            logging.info("Encoded %s video frames (target: %s)", frame_count, target_frames)
 
         # Decode and re-encode audio frames
         if audio_stream:
@@ -353,7 +304,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
             for packet in audio_stream.encode():
                 output_container.mux(packet)
 
-            logging.info(f"Encoded {audio_frame_count} audio frames")
+            logging.info("Encoded %s audio frames", audio_frame_count)
 
         # Close containers
         output_container.close()
@@ -380,7 +331,7 @@ def parse_width_height_from_res(resolution: str):
         "1:1 (1152 x 1152)": {"width": 1152, "height": 1152},
         "4:3 (1536 x 1152)": {"width": 1536, "height": 1152},
         "3:4 (1152 x 1536)": {"width": 1152, "height": 1536},
-        "21:9 (2560 x 1080)": {"width": 2560, "height": 1080},
+        # "21:9 (2560 x 1080)": {"width": 2560, "height": 1080},
     }
     return res_map.get(resolution, {"width": 1920, "height": 1080})
 
@@ -433,11 +384,11 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
                     "negative_prompt",
                     multiline=True,
                     default="<synthetic> <scene cut> gopro, bright, contrast, static, overexposed, vignette, "
-                            "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
-                            "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
-                            "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
-                            "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
-                            "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
+                    "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
+                    "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
+                    "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
+                    "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
+                    "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
                     tooltip="Negative prompt text",
                 ),
                 comfy_io.Combo.Input(
@@ -448,14 +399,14 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
                         "1:1 (1152 x 1152)",
                         "4:3 (1536 x 1152)",
                         "3:4 (1152 x 1536)",
-                        "21:9 (2560 x 1080)",
+                        # "21:9 (2560 x 1080)",
                     ],
                     default="16:9 (1920 x 1080)",
                     tooltip="Resolution of the output video",
                 ),
                 comfy_io.Float.Input(
                     "prompt_adherence",
-                    default=10.0,
+                    default=4.5,
                     min=1.0,
                     max=20.0,
                     step=1.0,
@@ -469,10 +420,11 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
                     step=1,
                     display_mode=comfy_io.NumberDisplay.number,
                     tooltip="Random seed value",
+                    control_after_generate=True,
                 ),
                 comfy_io.Int.Input(
                     "steps",
-                    default=100,
+                    default=33,
                     min=1,
                     max=100,
                     step=1,
@@ -499,7 +451,7 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
         seed: int,
         steps: int,
     ) -> comfy_io.NodeOutput:
-        validate_input_image(image, True)
+        validate_image_dimensions(image, min_width=300, min_height=300, max_height=MAX_HEIGHT, max_width=MAX_WIDTH)
         validate_prompts(prompt, negative_prompt, MOONVALLEY_MAREY_MAX_PROMPT_LENGTH)
         width_height = parse_width_height_from_res(resolution)
 
@@ -513,12 +465,11 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
             steps=steps,
             seed=seed,
             guidance_scale=prompt_adherence,
-            num_frames=128,
             width=width_height["width"],
             height=width_height["height"],
             use_negative_prompts=True,
         )
-        """Upload image to comfy backend to have a URL available for further processing"""
+
         # Get MIME type from tensor - assuming PNG format for image tensors
         mime_type = "image/png"
 
@@ -571,11 +522,11 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
                     "negative_prompt",
                     multiline=True,
                     default="<synthetic> <scene cut> gopro, bright, contrast, static, overexposed, vignette, "
-                            "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
-                            "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
-                            "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
-                            "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
-                            "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
+                    "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
+                    "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
+                    "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
+                    "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
+                    "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
                     tooltip="Negative prompt text",
                 ),
                 comfy_io.Int.Input(
@@ -591,7 +542,7 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
                 comfy_io.Video.Input(
                     "video",
                     tooltip="The reference video used to generate the output video. Must be at least 5 seconds long. "
-                            "Videos longer than 5s will be automatically trimmed. Only MP4 format supported.",
+                    "Videos longer than 5s will be automatically trimmed. Only MP4 format supported.",
                 ),
                 comfy_io.Combo.Input(
                     "control_type",
@@ -608,6 +559,15 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
                     tooltip="Only used if control_type is 'Motion Transfer'",
                     optional=True,
                 ),
+                comfy_io.Int.Input(
+                    "steps",
+                    default=33,
+                    min=1,
+                    max=100,
+                    step=1,
+                    display_mode=comfy_io.NumberDisplay.number,
+                    tooltip="Number of inference steps",
+                ),
             ],
             outputs=[comfy_io.Video.Output()],
             hidden=[
@@ -627,6 +587,8 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
         video: Optional[VideoInput] = None,
         control_type: str = "Motion Transfer",
         motion_intensity: Optional[int] = 100,
+        steps=33,
+        prompt_adherence=4.5,
     ) -> comfy_io.NodeOutput:
         auth = {
             "auth_token": cls.hidden.auth_token_comfy_org,
@@ -636,7 +598,6 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
         validated_video = validate_video_to_video_input(video)
         video_url = await upload_video_to_comfyapi(validated_video, auth_kwargs=auth)
 
-        """Validate prompts and inference input"""
         validate_prompts(prompt, negative_prompt)
 
         # Only include motion_intensity for Motion Transfer
@@ -648,6 +609,8 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
             negative_prompt=negative_prompt,
             seed=seed,
             control_params=control_params,
+            steps=steps,
+            guidance_scale=prompt_adherence,
         )
 
         control = parse_control_parameter(control_type)
@@ -699,11 +662,11 @@ class MoonvalleyTxt2VideoNode(comfy_io.ComfyNode):
                     "negative_prompt",
                     multiline=True,
                     default="<synthetic> <scene cut> gopro, bright, contrast, static, overexposed, vignette, "
-                            "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
-                            "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
-                            "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
-                            "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
-                            "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
+                    "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
+                    "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
+                    "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
+                    "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
+                    "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
                     tooltip="Negative prompt text",
                 ),
                 comfy_io.Combo.Input(
@@ -721,7 +684,7 @@ class MoonvalleyTxt2VideoNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Float.Input(
                     "prompt_adherence",
-                    default=10.0,
+                    default=4.0,
                     min=1.0,
                     max=20.0,
                     step=1.0,
@@ -734,11 +697,12 @@ class MoonvalleyTxt2VideoNode(comfy_io.ComfyNode):
                     max=4294967295,
                     step=1,
                     display_mode=comfy_io.NumberDisplay.number,
+                    control_after_generate=True,
                     tooltip="Random seed value",
                 ),
                 comfy_io.Int.Input(
                     "steps",
-                    default=100,
+                    default=33,
                     min=1,
                     max=100,
                     step=1,

comfy_api_nodes/nodes_pixverse.py

@@ -1,5 +1,7 @@
 from inspect import cleandoc
 from typing import Optional
+from typing_extensions import override
+from io import BytesIO
 from comfy_api_nodes.apis.pixverse_api import (
     PixverseTextVideoRequest,
     PixverseImageVideoRequest,
@@ -26,12 +28,11 @@ from comfy_api_nodes.apinode_utils import (
     tensor_to_bytesio,
     validate_string,
 )
-from comfy.comfy_types.node_typing import IO, ComfyNodeABC
 from comfy_api.input_impl import VideoFromFile
+from comfy_api.latest import ComfyExtension, io as comfy_io
 
 import torch
 import aiohttp
-from io import BytesIO
 
 
 AVERAGE_DURATION_T2V = 32
@@ -72,100 +73,101 @@ async def upload_image_to_pixverse(image: torch.Tensor, auth_kwargs=None):
     return response_upload.Resp.img_id
 
 
-class PixverseTemplateNode:
+class PixverseTemplateNode(comfy_io.ComfyNode):
     """
     Select template for PixVerse Video generation.
     """
 
-    RETURN_TYPES = (PixverseIO.TEMPLATE,)
-    RETURN_NAMES = ("pixverse_template",)
-    FUNCTION = "create_template"
-    CATEGORY = "api node/video/PixVerse"
+    @classmethod
+    def define_schema(cls) -> comfy_io.Schema:
+        return comfy_io.Schema(
+            node_id="PixverseTemplateNode",
+            display_name="PixVerse Template",
+            category="api node/video/PixVerse",
+            inputs=[
+                comfy_io.Combo.Input("template", options=list(pixverse_templates.keys())),
+            ],
+            outputs=[comfy_io.Custom(PixverseIO.TEMPLATE).Output(display_name="pixverse_template")],
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "template": (list(pixverse_templates.keys()),),
-            }
-        }
-
-    def create_template(self, template: str):
+    def execute(cls, template: str) -> comfy_io.NodeOutput:
         template_id = pixverse_templates.get(template, None)
         if template_id is None:
             raise Exception(f"Template '{template}' is not recognized.")
         # just return the integer
-        return (template_id,)
+        return comfy_io.NodeOutput(template_id)
 
 
-class PixverseTextToVideoNode(ComfyNodeABC):
+class PixverseTextToVideoNode(comfy_io.ComfyNode):
     """
     Generates videos based on prompt and output_size.
     """
 
-    RETURN_TYPES = (IO.VIDEO,)
-    DESCRIPTION = cleandoc(__doc__ or "")  # Handle potential None value
-    FUNCTION = "api_call"
-    API_NODE = True
-    CATEGORY = "api node/video/PixVerse"
+    @classmethod
+    def define_schema(cls) -> comfy_io.Schema:
+        return comfy_io.Schema(
+            node_id="PixverseTextToVideoNode",
+            display_name="PixVerse Text to Video",
+            category="api node/video/PixVerse",
+            description=cleandoc(cls.__doc__ or ""),
+            inputs=[
+                comfy_io.String.Input(
+                    "prompt",
+                    multiline=True,
+                    default="",
+                    tooltip="Prompt for the video generation",
+                ),
+                comfy_io.Combo.Input(
+                    "aspect_ratio",
+                    options=PixverseAspectRatio,
+                ),
+                comfy_io.Combo.Input(
+                    "quality",
+                    options=PixverseQuality,
+                    default=PixverseQuality.res_540p,
+                ),
+                comfy_io.Combo.Input(
+                    "duration_seconds",
+                    options=PixverseDuration,
+                ),
+                comfy_io.Combo.Input(
+                    "motion_mode",
+                    options=PixverseMotionMode,
+                ),
+                comfy_io.Int.Input(
+                    "seed",
+                    default=0,
+                    min=0,
+                    max=2147483647,
+                    control_after_generate=True,
+                    tooltip="Seed for video generation.",
+                ),
+                comfy_io.String.Input(
+                    "negative_prompt",
+                    default="",
+                    multiline=True,
+                    tooltip="An optional text description of undesired elements on an image.",
+                    optional=True,
+                ),
+                comfy_io.Custom(PixverseIO.TEMPLATE).Input(
+                    "pixverse_template",
+                    tooltip="An optional template to influence style of generation, created by the PixVerse Template node.",
+                    optional=True,
+                ),
+            ],
+            outputs=[comfy_io.Video.Output()],
+            hidden=[
+                comfy_io.Hidden.auth_token_comfy_org,
+                comfy_io.Hidden.api_key_comfy_org,
+                comfy_io.Hidden.unique_id,
+            ],
+            is_api_node=True,
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "prompt": (
-                    IO.STRING,
-                    {
-                        "multiline": True,
-                        "default": "",
-                        "tooltip": "Prompt for the video generation",
-                    },
-                ),
-                "aspect_ratio": ([ratio.value for ratio in PixverseAspectRatio],),
-                "quality": (
-                    [resolution.value for resolution in PixverseQuality],
-                    {
-                        "default": PixverseQuality.res_540p,
-                    },
-                ),
-                "duration_seconds": ([dur.value for dur in PixverseDuration],),
-                "motion_mode": ([mode.value for mode in PixverseMotionMode],),
-                "seed": (
-                    IO.INT,
-                    {
-                        "default": 0,
-                        "min": 0,
-                        "max": 2147483647,
-                        "control_after_generate": True,
-                        "tooltip": "Seed for video generation.",
-                    },
-                ),
-            },
-            "optional": {
-                "negative_prompt": (
-                    IO.STRING,
-                    {
-                        "default": "",
-                        "forceInput": True,
-                        "tooltip": "An optional text description of undesired elements on an image.",
-                    },
-                ),
-                "pixverse_template": (
-                    PixverseIO.TEMPLATE,
-                    {
-                        "tooltip": "An optional template to influence style of generation, created by the PixVerse Template node."
-                    },
-                ),
-            },
-            "hidden": {
-                "auth_token": "AUTH_TOKEN_COMFY_ORG",
-                "comfy_api_key": "API_KEY_COMFY_ORG",
-                "unique_id": "UNIQUE_ID",
-            },
-        }
-
-    async def api_call(
-        self,
+    async def execute(
+        cls,
         prompt: str,
         aspect_ratio: str,
         quality: str,
@@ -174,9 +176,7 @@ class PixverseTextToVideoNode(ComfyNodeABC):
         seed,
         negative_prompt: str = None,
         pixverse_template: int = None,
-        unique_id: Optional[str] = None,
-        **kwargs,
-    ):
+    ) -> comfy_io.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         # 1080p is limited to 5 seconds duration
         # only normal motion_mode supported for 1080p or for non-5 second duration
@@ -186,6 +186,10 @@ class PixverseTextToVideoNode(ComfyNodeABC):
         elif duration_seconds != PixverseDuration.dur_5:
             motion_mode = PixverseMotionMode.normal
 
+        auth = {
+            "auth_token": cls.hidden.auth_token_comfy_org,
+            "comfy_api_key": cls.hidden.api_key_comfy_org,
+        }
         operation = SynchronousOperation(
             endpoint=ApiEndpoint(
                 path="/proxy/pixverse/video/text/generate",
@@ -203,7 +207,7 @@ class PixverseTextToVideoNode(ComfyNodeABC):
                 template_id=pixverse_template,
                 seed=seed,
             ),
-            auth_kwargs=kwargs,
+            auth_kwargs=auth,
         )
         response_api = await operation.execute()
 
@@ -224,8 +228,8 @@ class PixverseTextToVideoNode(ComfyNodeABC):
                 PixverseStatus.deleted,
             ],
             status_extractor=lambda x: x.Resp.status,
-            auth_kwargs=kwargs,
-            node_id=unique_id,
+            auth_kwargs=auth,
+            node_id=cls.hidden.unique_id,
             result_url_extractor=get_video_url_from_response,
             estimated_duration=AVERAGE_DURATION_T2V,
         )
@@ -233,77 +237,75 @@ class PixverseTextToVideoNode(ComfyNodeABC):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.Resp.url) as vid_response:
-                return (VideoFromFile(BytesIO(await vid_response.content.read())),)
+                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
 
-class PixverseImageToVideoNode(ComfyNodeABC):
+class PixverseImageToVideoNode(comfy_io.ComfyNode):
     """
     Generates videos based on prompt and output_size.
     """
 
-    RETURN_TYPES = (IO.VIDEO,)
-    DESCRIPTION = cleandoc(__doc__ or "")  # Handle potential None value
-    FUNCTION = "api_call"
-    API_NODE = True
-    CATEGORY = "api node/video/PixVerse"
+    @classmethod
+    def define_schema(cls) -> comfy_io.Schema:
+        return comfy_io.Schema(
+            node_id="PixverseImageToVideoNode",
+            display_name="PixVerse Image to Video",
+            category="api node/video/PixVerse",
+            description=cleandoc(cls.__doc__ or ""),
+            inputs=[
+                comfy_io.Image.Input("image"),
+                comfy_io.String.Input(
+                    "prompt",
+                    multiline=True,
+                    default="",
+                    tooltip="Prompt for the video generation",
+                ),
+                comfy_io.Combo.Input(
+                    "quality",
+                    options=PixverseQuality,
+                    default=PixverseQuality.res_540p,
+                ),
+                comfy_io.Combo.Input(
+                    "duration_seconds",
+                    options=PixverseDuration,
+                ),
+                comfy_io.Combo.Input(
+                    "motion_mode",
+                    options=PixverseMotionMode,
+                ),
+                comfy_io.Int.Input(
+                    "seed",
+                    default=0,
+                    min=0,
+                    max=2147483647,
+                    control_after_generate=True,
+                    tooltip="Seed for video generation.",
+                ),
+                comfy_io.String.Input(
+                    "negative_prompt",
+                    default="",
+                    multiline=True,
+                    tooltip="An optional text description of undesired elements on an image.",
+                    optional=True,
+                ),
+                comfy_io.Custom(PixverseIO.TEMPLATE).Input(
+                    "pixverse_template",
+                    tooltip="An optional template to influence style of generation, created by the PixVerse Template node.",
+                    optional=True,
+                ),
+            ],
+            outputs=[comfy_io.Video.Output()],
+            hidden=[
+                comfy_io.Hidden.auth_token_comfy_org,
+                comfy_io.Hidden.api_key_comfy_org,
+                comfy_io.Hidden.unique_id,
+            ],
+            is_api_node=True,
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "image": (IO.IMAGE,),
-                "prompt": (
-                    IO.STRING,
-                    {
-                        "multiline": True,
-                        "default": "",
-                        "tooltip": "Prompt for the video generation",
-                    },
-                ),
-                "quality": (
-                    [resolution.value for resolution in PixverseQuality],
-                    {
-                        "default": PixverseQuality.res_540p,
-                    },
-                ),
-                "duration_seconds": ([dur.value for dur in PixverseDuration],),
-                "motion_mode": ([mode.value for mode in PixverseMotionMode],),
-                "seed": (
-                    IO.INT,
-                    {
-                        "default": 0,
-                        "min": 0,
-                        "max": 2147483647,
-                        "control_after_generate": True,
-                        "tooltip": "Seed for video generation.",
-                    },
-                ),
-            },
-            "optional": {
-                "negative_prompt": (
-                    IO.STRING,
-                    {
-                        "default": "",
-                        "forceInput": True,
-                        "tooltip": "An optional text description of undesired elements on an image.",
-                    },
-                ),
-                "pixverse_template": (
-                    PixverseIO.TEMPLATE,
-                    {
-                        "tooltip": "An optional template to influence style of generation, created by the PixVerse Template node."
-                    },
-                ),
-            },
-            "hidden": {
-                "auth_token": "AUTH_TOKEN_COMFY_ORG",
-                "comfy_api_key": "API_KEY_COMFY_ORG",
-                "unique_id": "UNIQUE_ID",
-            },
-        }
-
-    async def api_call(
-        self,
+    async def execute(
+        cls,
         image: torch.Tensor,
         prompt: str,
         quality: str,
@@ -312,11 +314,13 @@ class PixverseImageToVideoNode(ComfyNodeABC):
         seed,
         negative_prompt: str = None,
         pixverse_template: int = None,
-        unique_id: Optional[str] = None,
-        **kwargs,
-    ):
+    ) -> comfy_io.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
-        img_id = await upload_image_to_pixverse(image, auth_kwargs=kwargs)
+        auth = {
+            "auth_token": cls.hidden.auth_token_comfy_org,
+            "comfy_api_key": cls.hidden.api_key_comfy_org,
+        }
+        img_id = await upload_image_to_pixverse(image, auth_kwargs=auth)
 
         # 1080p is limited to 5 seconds duration
         # only normal motion_mode supported for 1080p or for non-5 second duration
@@ -343,7 +347,7 @@ class PixverseImageToVideoNode(ComfyNodeABC):
                 template_id=pixverse_template,
                 seed=seed,
             ),
-            auth_kwargs=kwargs,
+            auth_kwargs=auth,
         )
         response_api = await operation.execute()
 
@@ -364,8 +368,8 @@ class PixverseImageToVideoNode(ComfyNodeABC):
                 PixverseStatus.deleted,
             ],
             status_extractor=lambda x: x.Resp.status,
-            auth_kwargs=kwargs,
-            node_id=unique_id,
+            auth_kwargs=auth,
+            node_id=cls.hidden.unique_id,
             result_url_extractor=get_video_url_from_response,
             estimated_duration=AVERAGE_DURATION_I2V,
         )
@@ -373,72 +377,71 @@ class PixverseImageToVideoNode(ComfyNodeABC):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.Resp.url) as vid_response:
-                return (VideoFromFile(BytesIO(await vid_response.content.read())),)
+                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
 
-class PixverseTransitionVideoNode(ComfyNodeABC):
+class PixverseTransitionVideoNode(comfy_io.ComfyNode):
     """
     Generates videos based on prompt and output_size.
     """
 
-    RETURN_TYPES = (IO.VIDEO,)
-    DESCRIPTION = cleandoc(__doc__ or "")  # Handle potential None value
-    FUNCTION = "api_call"
-    API_NODE = True
-    CATEGORY = "api node/video/PixVerse"
+    @classmethod
+    def define_schema(cls) -> comfy_io.Schema:
+        return comfy_io.Schema(
+            node_id="PixverseTransitionVideoNode",
+            display_name="PixVerse Transition Video",
+            category="api node/video/PixVerse",
+            description=cleandoc(cls.__doc__ or ""),
+            inputs=[
+                comfy_io.Image.Input("first_frame"),
+                comfy_io.Image.Input("last_frame"),
+                comfy_io.String.Input(
+                    "prompt",
+                    multiline=True,
+                    default="",
+                    tooltip="Prompt for the video generation",
+                ),
+                comfy_io.Combo.Input(
+                    "quality",
+                    options=PixverseQuality,
+                    default=PixverseQuality.res_540p,
+                ),
+                comfy_io.Combo.Input(
+                    "duration_seconds",
+                    options=PixverseDuration,
+                ),
+                comfy_io.Combo.Input(
+                    "motion_mode",
+                    options=PixverseMotionMode,
+                ),
+                comfy_io.Int.Input(
+                    "seed",
+                    default=0,
+                    min=0,
+                    max=2147483647,
+                    control_after_generate=True,
+                    tooltip="Seed for video generation.",
+                ),
+                comfy_io.String.Input(
+                    "negative_prompt",
+                    default="",
+                    multiline=True,
+                    tooltip="An optional text description of undesired elements on an image.",
+                    optional=True,
+                ),
+            ],
+            outputs=[comfy_io.Video.Output()],
+            hidden=[
+                comfy_io.Hidden.auth_token_comfy_org,
+                comfy_io.Hidden.api_key_comfy_org,
+                comfy_io.Hidden.unique_id,
+            ],
+            is_api_node=True,
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "first_frame": (IO.IMAGE,),
-                "last_frame": (IO.IMAGE,),
-                "prompt": (
-                    IO.STRING,
-                    {
-                        "multiline": True,
-                        "default": "",
-                        "tooltip": "Prompt for the video generation",
-                    },
-                ),
-                "quality": (
-                    [resolution.value for resolution in PixverseQuality],
-                    {
-                        "default": PixverseQuality.res_540p,
-                    },
-                ),
-                "duration_seconds": ([dur.value for dur in PixverseDuration],),
-                "motion_mode": ([mode.value for mode in PixverseMotionMode],),
-                "seed": (
-                    IO.INT,
-                    {
-                        "default": 0,
-                        "min": 0,
-                        "max": 2147483647,
-                        "control_after_generate": True,
-                        "tooltip": "Seed for video generation.",
-                    },
-                ),
-            },
-            "optional": {
-                "negative_prompt": (
-                    IO.STRING,
-                    {
-                        "default": "",
-                        "forceInput": True,
-                        "tooltip": "An optional text description of undesired elements on an image.",
-                    },
-                ),
-            },
-            "hidden": {
-                "auth_token": "AUTH_TOKEN_COMFY_ORG",
-                "comfy_api_key": "API_KEY_COMFY_ORG",
-                "unique_id": "UNIQUE_ID",
-            },
-        }
-
-    async def api_call(
-        self,
+    async def execute(
+        cls,
         first_frame: torch.Tensor,
         last_frame: torch.Tensor,
         prompt: str,
@@ -447,12 +450,14 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
         motion_mode: str,
         seed,
         negative_prompt: str = None,
-        unique_id: Optional[str] = None,
-        **kwargs,
-    ):
+    ) -> comfy_io.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
-        first_frame_id = await upload_image_to_pixverse(first_frame, auth_kwargs=kwargs)
-        last_frame_id = await upload_image_to_pixverse(last_frame, auth_kwargs=kwargs)
+        auth = {
+            "auth_token": cls.hidden.auth_token_comfy_org,
+            "comfy_api_key": cls.hidden.api_key_comfy_org,
+        }
+        first_frame_id = await upload_image_to_pixverse(first_frame, auth_kwargs=auth)
+        last_frame_id = await upload_image_to_pixverse(last_frame, auth_kwargs=auth)
 
         # 1080p is limited to 5 seconds duration
         # only normal motion_mode supported for 1080p or for non-5 second duration
@@ -479,7 +484,7 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
                 negative_prompt=negative_prompt if negative_prompt else None,
                 seed=seed,
             ),
-            auth_kwargs=kwargs,
+            auth_kwargs=auth,
         )
         response_api = await operation.execute()
 
@@ -500,8 +505,8 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
                 PixverseStatus.deleted,
             ],
             status_extractor=lambda x: x.Resp.status,
-            auth_kwargs=kwargs,
-            node_id=unique_id,
+            auth_kwargs=auth,
+            node_id=cls.hidden.unique_id,
             result_url_extractor=get_video_url_from_response,
             estimated_duration=AVERAGE_DURATION_T2V,
         )
@@ -509,19 +514,19 @@ class PixverseTransitionVideoNode(ComfyNodeABC):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.Resp.url) as vid_response:
-                return (VideoFromFile(BytesIO(await vid_response.content.read())),)
+                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
 
-NODE_CLASS_MAPPINGS = {
-    "PixverseTextToVideoNode": PixverseTextToVideoNode,
-    "PixverseImageToVideoNode": PixverseImageToVideoNode,
-    "PixverseTransitionVideoNode": PixverseTransitionVideoNode,
-    "PixverseTemplateNode": PixverseTemplateNode,
-}
+class PixVerseExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+        return [
+            PixverseTextToVideoNode,
+            PixverseImageToVideoNode,
+            PixverseTransitionVideoNode,
+            PixverseTemplateNode,
+        ]
 
-NODE_DISPLAY_NAME_MAPPINGS = {
-    "PixverseTextToVideoNode": "PixVerse Text to Video",
-    "PixverseImageToVideoNode": "PixVerse Image to Video",
-    "PixverseTransitionVideoNode": "PixVerse Transition Video",
-    "PixverseTemplateNode": "PixVerse Template",
-}
+
+async def comfy_entrypoint() -> PixVerseExtension:
+    return PixVerseExtension()

comfy_api_nodes/nodes_recraft.py

@@ -35,57 +35,64 @@ from server import PromptServer
 import torch
 from io import BytesIO
 from PIL import UnidentifiedImageError
+import aiohttp
 
 
 async def handle_recraft_file_request(
-        image: torch.Tensor,
-        path: str,
-        mask: torch.Tensor=None,
-        total_pixels=4096*4096,
-        timeout=1024,
-        request=None,
-        auth_kwargs: dict[str,str] = None,
-    ) -> list[BytesIO]:
-        """
-        Handle sending common Recraft file-only request to get back file bytes.
-        """
-        if request is None:
-            request = EmptyRequest()
-
-        files = {
-            'image': tensor_to_bytesio(image, total_pixels=total_pixels).read()
-        }
-        if mask is not None:
-            files['mask'] = tensor_to_bytesio(mask, total_pixels=total_pixels).read()
-
-        operation = SynchronousOperation(
-            endpoint=ApiEndpoint(
-                path=path,
-                method=HttpMethod.POST,
-                request_model=type(request),
-                response_model=RecraftImageGenerationResponse,
-            ),
-            request=request,
-            files=files,
-            content_type="multipart/form-data",
-            auth_kwargs=auth_kwargs,
-            multipart_parser=recraft_multipart_parser,
-        )
-        response: RecraftImageGenerationResponse = await operation.execute()
-        all_bytesio = []
-        if response.image is not None:
-            all_bytesio.append(await download_url_to_bytesio(response.image.url, timeout=timeout))
-        else:
-            for data in response.data:
-                all_bytesio.append(await download_url_to_bytesio(data.url, timeout=timeout))
-
-        return all_bytesio
-
-
-def recraft_multipart_parser(data, parent_key=None, formatter: callable=None, converted_to_check: list[list]=None, is_list=False) -> dict:
+    image: torch.Tensor,
+    path: str,
+    mask: torch.Tensor=None,
+    total_pixels=4096*4096,
+    timeout=1024,
+    request=None,
+    auth_kwargs: dict[str,str] = None,
+) -> list[BytesIO]:
     """
-    Formats data such that multipart/form-data will work with requests library
-    when both files and data are present.
+    Handle sending common Recraft file-only request to get back file bytes.
+    """
+    if request is None:
+        request = EmptyRequest()
+
+    files = {
+        'image': tensor_to_bytesio(image, total_pixels=total_pixels).read()
+    }
+    if mask is not None:
+        files['mask'] = tensor_to_bytesio(mask, total_pixels=total_pixels).read()
+
+    operation = SynchronousOperation(
+        endpoint=ApiEndpoint(
+            path=path,
+            method=HttpMethod.POST,
+            request_model=type(request),
+            response_model=RecraftImageGenerationResponse,
+        ),
+        request=request,
+        files=files,
+        content_type="multipart/form-data",
+        auth_kwargs=auth_kwargs,
+        multipart_parser=recraft_multipart_parser,
+    )
+    response: RecraftImageGenerationResponse = await operation.execute()
+    all_bytesio = []
+    if response.image is not None:
+        all_bytesio.append(await download_url_to_bytesio(response.image.url, timeout=timeout))
+    else:
+        for data in response.data:
+            all_bytesio.append(await download_url_to_bytesio(data.url, timeout=timeout))
+
+    return all_bytesio
+
+
+def recraft_multipart_parser(
+    data,
+    parent_key=None,
+    formatter: callable = None,
+    converted_to_check: list[list] = None,
+    is_list: bool = False,
+    return_mode: str = "formdata"  # "dict" | "formdata"
+) -> dict | aiohttp.FormData:
+    """
+    Formats data such that multipart/form-data will work with aiohttp library when both files and data are present.
 
     The OpenAI client that Recraft uses has a bizarre way of serializing lists:
 
@@ -103,23 +110,23 @@ def recraft_multipart_parser(data, parent_key=None, formatter: callable=None, co
     # Modification of a function that handled a different type of multipart parsing, big ups:
     # https://gist.github.com/kazqvaizer/4cebebe5db654a414132809f9f88067b
 
-    def handle_converted_lists(data, parent_key, lists_to_check=tuple[list]):
+    def handle_converted_lists(item, parent_key, lists_to_check=tuple[list]):
         # if list already exists exists, just extend list with data
         for check_list in lists_to_check:
             for conv_tuple in check_list:
-                if conv_tuple[0] == parent_key and type(conv_tuple[1]) is list:
-                    conv_tuple[1].append(formatter(data))
+                if conv_tuple[0] == parent_key and isinstance(conv_tuple[1], list):
+                    conv_tuple[1].append(formatter(item))
                     return True
         return False
 
     if converted_to_check is None:
         converted_to_check = []
 
-
+    effective_mode = return_mode if parent_key is None else "dict"
     if formatter is None:
         formatter = lambda v: v  # Multipart representation of value
 
-    if type(data) is not dict:
+    if not isinstance(data, dict):
         # if list already exists exists, just extend list with data
         added = handle_converted_lists(data, parent_key, converted_to_check)
         if added:
@@ -136,15 +143,24 @@ def recraft_multipart_parser(data, parent_key=None, formatter: callable=None, co
 
     for key, value in data.items():
         current_key = key if parent_key is None else f"{parent_key}[{key}]"
-        if type(value) is dict:
+        if isinstance(value, dict):
             converted.extend(recraft_multipart_parser(value, current_key, formatter, next_check).items())
-        elif type(value) is list:
+        elif isinstance(value, list):
             for ind, list_value in enumerate(value):
                 iter_key = f"{current_key}[]"
                 converted.extend(recraft_multipart_parser(list_value, iter_key, formatter, next_check, is_list=True).items())
         else:
             converted.append((current_key, formatter(value)))
 
+    if effective_mode == "formdata":
+        fd = aiohttp.FormData()
+        for k, v in dict(converted).items():
+            if isinstance(v, list):
+                for item in v:
+                    fd.add_field(k, str(item))
+            else:
+                fd.add_field(k, str(v))
+        return fd
     return dict(converted)
 
 

comfy_api_nodes/nodes_runway.py

@@ -200,11 +200,11 @@ class RunwayImageToVideoNodeGen3a(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "duration",
-                    options=[model.value for model in Duration],
+                    options=Duration,
                 ),
                 comfy_io.Combo.Input(
                     "ratio",
-                    options=[model.value for model in RunwayGen3aAspectRatio],
+                    options=RunwayGen3aAspectRatio,
                 ),
                 comfy_io.Int.Input(
                     "seed",
@@ -300,11 +300,11 @@ class RunwayImageToVideoNodeGen4(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "duration",
-                    options=[model.value for model in Duration],
+                    options=Duration,
                 ),
                 comfy_io.Combo.Input(
                     "ratio",
-                    options=[model.value for model in RunwayGen4TurboAspectRatio],
+                    options=RunwayGen4TurboAspectRatio,
                 ),
                 comfy_io.Int.Input(
                     "seed",
@@ -408,11 +408,11 @@ class RunwayFirstLastFrameNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "duration",
-                    options=[model.value for model in Duration],
+                    options=Duration,
                 ),
                 comfy_io.Combo.Input(
                     "ratio",
-                    options=[model.value for model in RunwayGen3aAspectRatio],
+                    options=RunwayGen3aAspectRatio,
                 ),
                 comfy_io.Int.Input(
                     "seed",

comfy_api_nodes/nodes_sora.py

@@ -0,0 +1,175 @@
+from typing import Optional
+from typing_extensions import override
+
+import torch
+from pydantic import BaseModel, Field
+from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api_nodes.apis.client import (
+    ApiEndpoint,
+    HttpMethod,
+    SynchronousOperation,
+    PollingOperation,
+    EmptyRequest,
+)
+from comfy_api_nodes.util.validation_utils import get_number_of_images
+
+from comfy_api_nodes.apinode_utils import (
+    download_url_to_video_output,
+    tensor_to_bytesio,
+)
+
+class Sora2GenerationRequest(BaseModel):
+    prompt: str = Field(...)
+    model: str = Field(...)
+    seconds: str = Field(...)
+    size: str = Field(...)
+
+
+class Sora2GenerationResponse(BaseModel):
+    id: str = Field(...)
+    error: Optional[dict] = Field(None)
+    status: Optional[str] = Field(None)
+
+
+class OpenAIVideoSora2(comfy_io.ComfyNode):
+    @classmethod
+    def define_schema(cls):
+        return comfy_io.Schema(
+            node_id="OpenAIVideoSora2",
+            display_name="OpenAI Sora - Video",
+            category="api node/video/Sora",
+            description="OpenAI video and audio generation.",
+            inputs=[
+                comfy_io.Combo.Input(
+                    "model",
+                    options=["sora-2", "sora-2-pro"],
+                    default="sora-2",
+                ),
+                comfy_io.String.Input(
+                    "prompt",
+                    multiline=True,
+                    default="",
+                    tooltip="Guiding text; may be empty if an input image is present.",
+                ),
+                comfy_io.Combo.Input(
+                    "size",
+                    options=[
+                        "720x1280",
+                        "1280x720",
+                        "1024x1792",
+                        "1792x1024",
+                    ],
+                    default="1280x720",
+                ),
+                comfy_io.Combo.Input(
+                    "duration",
+                    options=[4, 8, 12],
+                    default=8,
+                ),
+                comfy_io.Image.Input(
+                    "image",
+                    optional=True,
+                ),
+                comfy_io.Int.Input(
+                    "seed",
+                    default=0,
+                    min=0,
+                    max=2147483647,
+                    step=1,
+                    display_mode=comfy_io.NumberDisplay.number,
+                    control_after_generate=True,
+                    optional=True,
+                    tooltip="Seed to determine if node should re-run; "
+                            "actual results are nondeterministic regardless of seed.",
+                ),
+            ],
+            outputs=[
+                comfy_io.Video.Output(),
+            ],
+            hidden=[
+                comfy_io.Hidden.auth_token_comfy_org,
+                comfy_io.Hidden.api_key_comfy_org,
+                comfy_io.Hidden.unique_id,
+            ],
+            is_api_node=True,
+        )
+
+    @classmethod
+    async def execute(
+        cls,
+        model: str,
+        prompt: str,
+        size: str = "1280x720",
+        duration: int = 8,
+        seed: int = 0,
+        image: Optional[torch.Tensor] = None,
+    ):
+        if model == "sora-2" and size not in ("720x1280", "1280x720"):
+            raise ValueError("Invalid size for sora-2 model, only 720x1280 and 1280x720 are supported.")
+        files_input = None
+        if image is not None:
+            if get_number_of_images(image) != 1:
+                raise ValueError("Currently only one input image is supported.")
+            files_input = {"input_reference": ("image.png", tensor_to_bytesio(image), "image/png")}
+        auth = {
+            "auth_token": cls.hidden.auth_token_comfy_org,
+            "comfy_api_key": cls.hidden.api_key_comfy_org,
+        }
+        payload = Sora2GenerationRequest(
+            model=model,
+            prompt=prompt,
+            seconds=str(duration),
+            size=size,
+        )
+        initial_operation = SynchronousOperation(
+            endpoint=ApiEndpoint(
+                path="/proxy/openai/v1/videos",
+                method=HttpMethod.POST,
+                request_model=Sora2GenerationRequest,
+                response_model=Sora2GenerationResponse
+            ),
+            request=payload,
+            files=files_input,
+            auth_kwargs=auth,
+            content_type="multipart/form-data",
+        )
+        initial_response = await initial_operation.execute()
+        if initial_response.error:
+            raise Exception(initial_response.error.message)
+
+        model_time_multiplier = 1 if model == "sora-2" else 2
+        poll_operation = PollingOperation(
+            poll_endpoint=ApiEndpoint(
+                path=f"/proxy/openai/v1/videos/{initial_response.id}",
+                method=HttpMethod.GET,
+                request_model=EmptyRequest,
+                response_model=Sora2GenerationResponse
+            ),
+            completed_statuses=["completed"],
+            failed_statuses=["failed"],
+            status_extractor=lambda x: x.status,
+            auth_kwargs=auth,
+            poll_interval=8.0,
+            max_poll_attempts=160,
+            node_id=cls.hidden.unique_id,
+            estimated_duration=45 * (duration / 4) * model_time_multiplier,
+        )
+        await poll_operation.execute()
+        return comfy_io.NodeOutput(
+            await download_url_to_video_output(
+                f"/proxy/openai/v1/videos/{initial_response.id}/content",
+                auth_kwargs=auth,
+            )
+        )
+
+
+class OpenAISoraExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+        return [
+            OpenAIVideoSora2,
+        ]
+
+
+async def comfy_entrypoint() -> OpenAISoraExtension:
+    return OpenAISoraExtension()

comfy_api_nodes/nodes_stability.py

@@ -82,8 +82,8 @@ class StabilityStableImageUltraNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "aspect_ratio",
-                    options=[x.value for x in StabilityAspectRatio],
-                    default=StabilityAspectRatio.ratio_1_1.value,
+                    options=StabilityAspectRatio,
+                    default=StabilityAspectRatio.ratio_1_1,
                     tooltip="Aspect ratio of generated image.",
                 ),
                 comfy_io.Combo.Input(
@@ -217,12 +217,12 @@ class StabilityStableImageSD_3_5Node(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "model",
-                    options=[x.value for x in Stability_SD3_5_Model],
+                    options=Stability_SD3_5_Model,
                 ),
                 comfy_io.Combo.Input(
                     "aspect_ratio",
-                    options=[x.value for x in StabilityAspectRatio],
-                    default=StabilityAspectRatio.ratio_1_1.value,
+                    options=StabilityAspectRatio,
+                    default=StabilityAspectRatio.ratio_1_1,
                     tooltip="Aspect ratio of generated image.",
                 ),
                 comfy_io.Combo.Input(

comfy_api_nodes/nodes_veo2.py

@@ -215,7 +215,7 @@ class VeoVideoGenerationNode(comfy_io.ComfyNode):
         initial_response = await initial_operation.execute()
         operation_name = initial_response.name
 
-        logging.info(f"Veo generation started with operation name: {operation_name}")
+        logging.info("Veo generation started with operation name: %s", operation_name)
 
         # Define status extractor function
         def status_extractor(response):

comfy_api_nodes/nodes_vidu.py

@@ -173,8 +173,8 @@ class ViduTextToVideoNode(comfy_io.ComfyNode):
             inputs=[
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in VideoModelName],
-                    default=VideoModelName.vidu_q1.value,
+                    options=VideoModelName,
+                    default=VideoModelName.vidu_q1,
                     tooltip="Model name",
                 ),
                 comfy_io.String.Input(
@@ -205,22 +205,22 @@ class ViduTextToVideoNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "aspect_ratio",
-                    options=[model.value for model in AspectRatio],
-                    default=AspectRatio.r_16_9.value,
+                    options=AspectRatio,
+                    default=AspectRatio.r_16_9,
                     tooltip="The aspect ratio of the output video",
                     optional=True,
                 ),
                 comfy_io.Combo.Input(
                     "resolution",
-                    options=[model.value for model in Resolution],
-                    default=Resolution.r_1080p.value,
+                    options=Resolution,
+                    default=Resolution.r_1080p,
                     tooltip="Supported values may vary by model & duration",
                     optional=True,
                 ),
                 comfy_io.Combo.Input(
                     "movement_amplitude",
-                    options=[model.value for model in MovementAmplitude],
-                    default=MovementAmplitude.auto.value,
+                    options=MovementAmplitude,
+                    default=MovementAmplitude.auto,
                     tooltip="The movement amplitude of objects in the frame",
                     optional=True,
                 ),
@@ -278,8 +278,8 @@ class ViduImageToVideoNode(comfy_io.ComfyNode):
             inputs=[
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in VideoModelName],
-                    default=VideoModelName.vidu_q1.value,
+                    options=VideoModelName,
+                    default=VideoModelName.vidu_q1,
                     tooltip="Model name",
                 ),
                 comfy_io.Image.Input(
@@ -316,14 +316,14 @@ class ViduImageToVideoNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "resolution",
-                    options=[model.value for model in Resolution],
-                    default=Resolution.r_1080p.value,
+                    options=Resolution,
+                    default=Resolution.r_1080p,
                     tooltip="Supported values may vary by model & duration",
                     optional=True,
                 ),
                 comfy_io.Combo.Input(
                     "movement_amplitude",
-                    options=[model.value for model in MovementAmplitude],
+                    options=MovementAmplitude,
                     default=MovementAmplitude.auto.value,
                     tooltip="The movement amplitude of objects in the frame",
                     optional=True,
@@ -388,8 +388,8 @@ class ViduReferenceVideoNode(comfy_io.ComfyNode):
             inputs=[
                 comfy_io.Combo.Input(
                     "model",
-                    options=[model.value for model in VideoModelName],
-                    default=VideoModelName.vidu_q1.value,
+                    options=VideoModelName,
+                    default=VideoModelName.vidu_q1,
                     tooltip="Model name",
                 ),
                 comfy_io.Image.Input(
@@ -424,8 +424,8 @@ class ViduReferenceVideoNode(comfy_io.ComfyNode):
                 ),
                 comfy_io.Combo.Input(
                     "aspect_ratio",
-                    options=[model.value for model in AspectRatio],
-                    default=AspectRatio.r_16_9.value,
+                    options=AspectRatio,
+                    default=AspectRatio.r_16_9,
                     tooltip="The aspect ratio of the output video",
                     optional=True,
                 ),

comfy_extras/nodes_audio.py

@@ -360,7 +360,7 @@ class RecordAudio:
     def load(self, audio):
         audio_path = folder_paths.get_annotated_filepath(audio)
 
-        waveform, sample_rate = torchaudio.load(audio_path)
+        waveform, sample_rate = load(audio_path)
         audio = {"waveform": waveform.unsqueeze(0), "sample_rate": sample_rate}
         return (audio, )
 

comfy_extras/nodes_compositing.py

@@ -1,6 +1,9 @@
 import torch
 import comfy.utils
 from enum import Enum
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
+
 
 def resize_mask(mask, shape):
     return torch.nn.functional.interpolate(mask.reshape((-1, 1, mask.shape[-2], mask.shape[-1])), size=(shape[0], shape[1]), mode="bilinear").squeeze(1)
@@ -101,24 +104,28 @@ def porter_duff_composite(src_image: torch.Tensor, src_alpha: torch.Tensor, dst_
     return out_image, out_alpha
 
 
-class PorterDuffImageComposite:
+class PorterDuffImageComposite(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "source": ("IMAGE",),
-                "source_alpha": ("MASK",),
-                "destination": ("IMAGE",),
-                "destination_alpha": ("MASK",),
-                "mode": ([mode.name for mode in PorterDuffMode], {"default": PorterDuffMode.DST.name}),
-            },
-        }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="PorterDuffImageComposite",
+            display_name="Porter-Duff Image Composite",
+            category="mask/compositing",
+            inputs=[
+                io.Image.Input("source"),
+                io.Mask.Input("source_alpha"),
+                io.Image.Input("destination"),
+                io.Mask.Input("destination_alpha"),
+                io.Combo.Input("mode", options=[mode.name for mode in PorterDuffMode], default=PorterDuffMode.DST.name),
+            ],
+            outputs=[
+                io.Image.Output(),
+                io.Mask.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("IMAGE", "MASK")
-    FUNCTION = "composite"
-    CATEGORY = "mask/compositing"
-
-    def composite(self, source: torch.Tensor, source_alpha: torch.Tensor, destination: torch.Tensor, destination_alpha: torch.Tensor, mode):
+    @classmethod
+    def execute(cls, source: torch.Tensor, source_alpha: torch.Tensor, destination: torch.Tensor, destination_alpha: torch.Tensor, mode) -> io.NodeOutput:
         batch_size = min(len(source), len(source_alpha), len(destination), len(destination_alpha))
         out_images = []
         out_alphas = []
@@ -150,45 +157,48 @@ class PorterDuffImageComposite:
             out_images.append(out_image)
             out_alphas.append(out_alpha.squeeze(2))
 
-        result = (torch.stack(out_images), torch.stack(out_alphas))
-        return result
+        return io.NodeOutput(torch.stack(out_images), torch.stack(out_alphas))
 
 
-class SplitImageWithAlpha:
+class SplitImageWithAlpha(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-                "required": {
-                    "image": ("IMAGE",),
-                }
-        }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SplitImageWithAlpha",
+            display_name="Split Image with Alpha",
+            category="mask/compositing",
+            inputs=[
+                io.Image.Input("image"),
+            ],
+            outputs=[
+                io.Image.Output(),
+                io.Mask.Output(),
+            ],
+        )
 
-    CATEGORY = "mask/compositing"
-    RETURN_TYPES = ("IMAGE", "MASK")
-    FUNCTION = "split_image_with_alpha"
-
-    def split_image_with_alpha(self, image: torch.Tensor):
+    @classmethod
+    def execute(cls, image: torch.Tensor) -> io.NodeOutput:
         out_images = [i[:,:,:3] for i in image]
         out_alphas = [i[:,:,3] if i.shape[2] > 3 else torch.ones_like(i[:,:,0]) for i in image]
-        result = (torch.stack(out_images), 1.0 - torch.stack(out_alphas))
-        return result
+        return io.NodeOutput(torch.stack(out_images), 1.0 - torch.stack(out_alphas))
 
 
-class JoinImageWithAlpha:
+class JoinImageWithAlpha(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-                "required": {
-                    "image": ("IMAGE",),
-                    "alpha": ("MASK",),
-                }
-        }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="JoinImageWithAlpha",
+            display_name="Join Image with Alpha",
+            category="mask/compositing",
+            inputs=[
+                io.Image.Input("image"),
+                io.Mask.Input("alpha"),
+            ],
+            outputs=[io.Image.Output()],
+        )
 
-    CATEGORY = "mask/compositing"
-    RETURN_TYPES = ("IMAGE",)
-    FUNCTION = "join_image_with_alpha"
-
-    def join_image_with_alpha(self, image: torch.Tensor, alpha: torch.Tensor):
+    @classmethod
+    def execute(cls, image: torch.Tensor, alpha: torch.Tensor) -> io.NodeOutput:
         batch_size = min(len(image), len(alpha))
         out_images = []
 
@@ -196,19 +206,18 @@ class JoinImageWithAlpha:
         for i in range(batch_size):
            out_images.append(torch.cat((image[i][:,:,:3], alpha[i].unsqueeze(2)), dim=2))
 
-        result = (torch.stack(out_images),)
-        return result
+        return io.NodeOutput(torch.stack(out_images))
 
 
-NODE_CLASS_MAPPINGS = {
-    "PorterDuffImageComposite": PorterDuffImageComposite,
-    "SplitImageWithAlpha": SplitImageWithAlpha,
-    "JoinImageWithAlpha": JoinImageWithAlpha,
-}
+class CompositingExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            PorterDuffImageComposite,
+            SplitImageWithAlpha,
+            JoinImageWithAlpha,
+        ]
 
 
-NODE_DISPLAY_NAME_MAPPINGS = {
-    "PorterDuffImageComposite": "Porter-Duff Image Composite",
-    "SplitImageWithAlpha": "Split Image with Alpha",
-    "JoinImageWithAlpha": "Join Image with Alpha",
-}
+async def comfy_entrypoint() -> CompositingExtension:
+    return CompositingExtension()

comfy_extras/nodes_edit_model.py

@@ -1,26 +1,38 @@
 import node_helpers
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 
-class ReferenceLatent:
+class ReferenceLatent(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"conditioning": ("CONDITIONING", ),
-                            },
-                "optional": {"latent": ("LATENT", ),}
-               }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="ReferenceLatent",
+            category="advanced/conditioning/edit_models",
+            description="This node sets the guiding latent for an edit model. If the model supports it you can chain multiple to set multiple reference images.",
+            inputs=[
+                io.Conditioning.Input("conditioning"),
+                io.Latent.Input("latent", optional=True),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ]
+        )
 
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "append"
-
-    CATEGORY = "advanced/conditioning/edit_models"
-    DESCRIPTION = "This node sets the guiding latent for an edit model. If the model supports it you can chain multiple to set multiple reference images."
-
-    def append(self, conditioning, latent=None):
+    @classmethod
+    def execute(cls, conditioning, latent=None) -> io.NodeOutput:
         if latent is not None:
             conditioning = node_helpers.conditioning_set_values(conditioning, {"reference_latents": [latent["samples"]]}, append=True)
-        return (conditioning, )
+        return io.NodeOutput(conditioning)
 
 
-NODE_CLASS_MAPPINGS = {
-    "ReferenceLatent": ReferenceLatent,
-}
+class EditModelExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            ReferenceLatent,
+        ]
+
+
+def comfy_entrypoint() -> EditModelExtension:
+    return EditModelExtension()

comfy_extras/nodes_eps.py

@@ -1,4 +1,9 @@
-class EpsilonScaling:
+from typing_extensions import override
+
+from comfy_api.latest import ComfyExtension, io
+
+
+class EpsilonScaling(io.ComfyNode):
     """
     Implements the Epsilon Scaling method from 'Elucidating the Exposure Bias in Diffusion Models'
     (https://arxiv.org/abs/2308.15321v6).
@@ -8,26 +13,28 @@ class EpsilonScaling:
     recommended by the paper for its practicality and effectiveness.
     """
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "model": ("MODEL",),
-                "scaling_factor": ("FLOAT", {
-                    "default": 1.005,
-                    "min": 0.5,
-                    "max": 1.5,
-                    "step": 0.001,
-                    "display": "number"
-                }),
-            }
-        }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="Epsilon Scaling",
+            category="model_patches/unet",
+            inputs=[
+                io.Model.Input("model"),
+                io.Float.Input(
+                    "scaling_factor",
+                    default=1.005,
+                    min=0.5,
+                    max=1.5,
+                    step=0.001,
+                    display_mode=io.NumberDisplay.number,
+                ),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "patch"
-
-    CATEGORY = "model_patches/unet"
-
-    def patch(self, model, scaling_factor):
+    @classmethod
+    def execute(cls, model, scaling_factor) -> io.NodeOutput:
         # Prevent division by zero, though the UI's min value should prevent this.
         if scaling_factor == 0:
             scaling_factor = 1e-9
@@ -53,8 +60,15 @@ class EpsilonScaling:
 
         model_clone.set_model_sampler_post_cfg_function(epsilon_scaling_function)
 
-        return (model_clone,)
+        return io.NodeOutput(model_clone)
 
-NODE_CLASS_MAPPINGS = {
-    "Epsilon Scaling": EpsilonScaling
-}
+
+class EpsilonScalingExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            EpsilonScaling,
+        ]
+
+async def comfy_entrypoint() -> EpsilonScalingExtension:
+    return EpsilonScalingExtension()

comfy_extras/nodes_flux.py

@@ -1,60 +1,80 @@
 import node_helpers
 import comfy.utils
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
-class CLIPTextEncodeFlux:
+
+class CLIPTextEncodeFlux(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "clip": ("CLIP", ),
-            "clip_l": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "t5xxl": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "guidance": ("FLOAT", {"default": 3.5, "min": 0.0, "max": 100.0, "step": 0.1}),
-            }}
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "encode"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="CLIPTextEncodeFlux",
+            category="advanced/conditioning/flux",
+            inputs=[
+                io.Clip.Input("clip"),
+                io.String.Input("clip_l", multiline=True, dynamic_prompts=True),
+                io.String.Input("t5xxl", multiline=True, dynamic_prompts=True),
+                io.Float.Input("guidance", default=3.5, min=0.0, max=100.0, step=0.1),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/conditioning/flux"
-
-    def encode(self, clip, clip_l, t5xxl, guidance):
+    @classmethod
+    def execute(cls, clip, clip_l, t5xxl, guidance) -> io.NodeOutput:
         tokens = clip.tokenize(clip_l)
         tokens["t5xxl"] = clip.tokenize(t5xxl)["t5xxl"]
 
-        return (clip.encode_from_tokens_scheduled(tokens, add_dict={"guidance": guidance}), )
+        return io.NodeOutput(clip.encode_from_tokens_scheduled(tokens, add_dict={"guidance": guidance}))
 
-class FluxGuidance:
+    encode = execute  # TODO: remove
+
+
+class FluxGuidance(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "conditioning": ("CONDITIONING", ),
-            "guidance": ("FLOAT", {"default": 3.5, "min": 0.0, "max": 100.0, "step": 0.1}),
-            }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxGuidance",
+            category="advanced/conditioning/flux",
+            inputs=[
+                io.Conditioning.Input("conditioning"),
+                io.Float.Input("guidance", default=3.5, min=0.0, max=100.0, step=0.1),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "append"
-
-    CATEGORY = "advanced/conditioning/flux"
-
-    def append(self, conditioning, guidance):
+    @classmethod
+    def execute(cls, conditioning, guidance) -> io.NodeOutput:
         c = node_helpers.conditioning_set_values(conditioning, {"guidance": guidance})
-        return (c, )
+        return io.NodeOutput(c)
+
+    append = execute  # TODO: remove
 
 
-class FluxDisableGuidance:
+class FluxDisableGuidance(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "conditioning": ("CONDITIONING", ),
-            }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxDisableGuidance",
+            category="advanced/conditioning/flux",
+            description="This node completely disables the guidance embed on Flux and Flux like models",
+            inputs=[
+                io.Conditioning.Input("conditioning"),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "append"
-
-    CATEGORY = "advanced/conditioning/flux"
-    DESCRIPTION = "This node completely disables the guidance embed on Flux and Flux like models"
-
-    def append(self, conditioning):
+    @classmethod
+    def execute(cls, conditioning) -> io.NodeOutput:
         c = node_helpers.conditioning_set_values(conditioning, {"guidance": None})
-        return (c, )
+        return io.NodeOutput(c)
+
+    append = execute  # TODO: remove
 
 
 PREFERED_KONTEXT_RESOLUTIONS = [
@@ -78,52 +98,73 @@ PREFERED_KONTEXT_RESOLUTIONS = [
 ]
 
 
-class FluxKontextImageScale:
+class FluxKontextImageScale(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"image": ("IMAGE", ),
-                            },
-               }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxKontextImageScale",
+            category="advanced/conditioning/flux",
+            description="This node resizes the image to one that is more optimal for flux kontext.",
+            inputs=[
+                io.Image.Input("image"),
+            ],
+            outputs=[
+                io.Image.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("IMAGE",)
-    FUNCTION = "scale"
-
-    CATEGORY = "advanced/conditioning/flux"
-    DESCRIPTION = "This node resizes the image to one that is more optimal for flux kontext."
-
-    def scale(self, image):
+    @classmethod
+    def execute(cls, image) -> io.NodeOutput:
         width = image.shape[2]
         height = image.shape[1]
         aspect_ratio = width / height
         _, width, height = min((abs(aspect_ratio - w / h), w, h) for w, h in PREFERED_KONTEXT_RESOLUTIONS)
         image = comfy.utils.common_upscale(image.movedim(-1, 1), width, height, "lanczos", "center").movedim(1, -1)
-        return (image, )
+        return io.NodeOutput(image)
+
+    scale = execute  # TODO: remove
 
 
-class FluxKontextMultiReferenceLatentMethod:
+class FluxKontextMultiReferenceLatentMethod(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "conditioning": ("CONDITIONING", ),
-            "reference_latents_method": (("offset", "index", "uxo/uno"), ),
-            }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxKontextMultiReferenceLatentMethod",
+            category="advanced/conditioning/flux",
+            inputs=[
+                io.Conditioning.Input("conditioning"),
+                io.Combo.Input(
+                    "reference_latents_method",
+                    options=["offset", "index", "uxo/uno"],
+                ),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+            is_experimental=True,
+        )
 
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "append"
-    EXPERIMENTAL = True
-
-    CATEGORY = "advanced/conditioning/flux"
-
-    def append(self, conditioning, reference_latents_method):
+    @classmethod
+    def execute(cls, conditioning, reference_latents_method) -> io.NodeOutput:
         if "uxo" in reference_latents_method or "uso" in reference_latents_method:
             reference_latents_method = "uxo"
         c = node_helpers.conditioning_set_values(conditioning, {"reference_latents_method": reference_latents_method})
-        return (c, )
+        return io.NodeOutput(c)
 
-NODE_CLASS_MAPPINGS = {
-    "CLIPTextEncodeFlux": CLIPTextEncodeFlux,
-    "FluxGuidance": FluxGuidance,
-    "FluxDisableGuidance": FluxDisableGuidance,
-    "FluxKontextImageScale": FluxKontextImageScale,
-    "FluxKontextMultiReferenceLatentMethod": FluxKontextMultiReferenceLatentMethod,
-}
+    append = execute  # TODO: remove
+
+
+class FluxExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            CLIPTextEncodeFlux,
+            FluxGuidance,
+            FluxDisableGuidance,
+            FluxKontextImageScale,
+            FluxKontextMultiReferenceLatentMethod,
+        ]
+
+
+async def comfy_entrypoint() -> FluxExtension:
+    return FluxExtension()

comfy_extras/nodes_latent.py

@@ -2,6 +2,8 @@ import comfy.utils
 import comfy_extras.nodes_post_processing
 import torch
 import nodes
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 
 def reshape_latent_to(target_shape, latent, repeat_batch=True):
@@ -13,17 +15,23 @@ def reshape_latent_to(target_shape, latent, repeat_batch=True):
         return latent
 
 
-class LatentAdd:
+class LatentAdd(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",)}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentAdd",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2):
+    @classmethod
+    def execute(cls, samples1, samples2) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -31,19 +39,25 @@ class LatentAdd:
 
         s2 = reshape_latent_to(s1.shape, s2)
         samples_out["samples"] = s1 + s2
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentSubtract:
+class LatentSubtract(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",)}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentSubtract",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2):
+    @classmethod
+    def execute(cls, samples1, samples2) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -51,41 +65,49 @@ class LatentSubtract:
 
         s2 = reshape_latent_to(s1.shape, s2)
         samples_out["samples"] = s1 - s2
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentMultiply:
+class LatentMultiply(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples": ("LATENT",),
-                              "multiplier": ("FLOAT", {"default": 1.0, "min": -10.0, "max": 10.0, "step": 0.01}),
-                             }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentMultiply",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples"),
+                io.Float.Input("multiplier", default=1.0, min=-10.0, max=10.0, step=0.01),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples, multiplier):
+    @classmethod
+    def execute(cls, samples, multiplier) -> io.NodeOutput:
         samples_out = samples.copy()
 
         s1 = samples["samples"]
         samples_out["samples"] = s1 * multiplier
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentInterpolate:
+class LatentInterpolate(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",),
-                              "samples2": ("LATENT",),
-                              "ratio": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}),
-                              }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentInterpolate",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+                io.Float.Input("ratio", default=1.0, min=0.0, max=1.0, step=0.01),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2, ratio):
+    @classmethod
+    def execute(cls, samples1, samples2, ratio) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -104,19 +126,26 @@ class LatentInterpolate:
         st = torch.nan_to_num(t / mt)
 
         samples_out["samples"] = st * (m1 * ratio + m2 * (1.0 - ratio))
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentConcat:
+class LatentConcat(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",), "dim": (["x", "-x", "y", "-y", "t", "-t"], )}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentConcat",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+                io.Combo.Input("dim", options=["x", "-x", "y", "-y", "t", "-t"]),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2, dim):
+    @classmethod
+    def execute(cls, samples1, samples2, dim) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -136,22 +165,27 @@ class LatentConcat:
             dim = -3
 
         samples_out["samples"] = torch.cat(c, dim=dim)
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentCut:
+class LatentCut(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"samples": ("LATENT",),
-                             "dim": (["x", "y", "t"], ),
-                             "index": ("INT", {"default": 0, "min": -nodes.MAX_RESOLUTION, "max": nodes.MAX_RESOLUTION, "step": 1}),
-                             "amount": ("INT", {"default": 1, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 1})}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentCut",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples"),
+                io.Combo.Input("dim", options=["x", "y", "t"]),
+                io.Int.Input("index", default=0, min=-nodes.MAX_RESOLUTION, max=nodes.MAX_RESOLUTION, step=1),
+                io.Int.Input("amount", default=1, min=1, max=nodes.MAX_RESOLUTION, step=1),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples, dim, index, amount):
+    @classmethod
+    def execute(cls, samples, dim, index, amount) -> io.NodeOutput:
         samples_out = samples.copy()
 
         s1 = samples["samples"]
@@ -171,19 +205,25 @@ class LatentCut:
             amount = min(-index, amount)
 
         samples_out["samples"] = torch.narrow(s1, dim, index, amount)
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentBatch:
+class LatentBatch(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",)}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentBatch",
+            category="latent/batch",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "batch"
-
-    CATEGORY = "latent/batch"
-
-    def batch(self, samples1, samples2):
+    @classmethod
+    def execute(cls, samples1, samples2) -> io.NodeOutput:
         samples_out = samples1.copy()
         s1 = samples1["samples"]
         s2 = samples2["samples"]
@@ -192,20 +232,25 @@ class LatentBatch:
         s = torch.cat((s1, s2), dim=0)
         samples_out["samples"] = s
         samples_out["batch_index"] = samples1.get("batch_index", [x for x in range(0, s1.shape[0])]) + samples2.get("batch_index", [x for x in range(0, s2.shape[0])])
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentBatchSeedBehavior:
+class LatentBatchSeedBehavior(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples": ("LATENT",),
-                              "seed_behavior": (["random", "fixed"],{"default": "fixed"}),}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentBatchSeedBehavior",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples"),
+                io.Combo.Input("seed_behavior", options=["random", "fixed"], default="fixed"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples, seed_behavior):
+    @classmethod
+    def execute(cls, samples, seed_behavior) -> io.NodeOutput:
         samples_out = samples.copy()
         latent = samples["samples"]
         if seed_behavior == "random":
@@ -215,41 +260,50 @@ class LatentBatchSeedBehavior:
             batch_number = samples_out.get("batch_index", [0])[0]
             samples_out["batch_index"] = [batch_number] * latent.shape[0]
 
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentApplyOperation:
+class LatentApplyOperation(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples": ("LATENT",),
-                             "operation": ("LATENT_OPERATION",),
-                             }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentApplyOperation",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Latent.Input("samples"),
+                io.LatentOperation.Input("operation"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def op(self, samples, operation):
+    @classmethod
+    def execute(cls, samples, operation) -> io.NodeOutput:
         samples_out = samples.copy()
 
         s1 = samples["samples"]
         samples_out["samples"] = operation(latent=s1)
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentApplyOperationCFG:
+class LatentApplyOperationCFG(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model": ("MODEL",),
-                             "operation": ("LATENT_OPERATION",),
-                              }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "patch"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentApplyOperationCFG",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Model.Input("model"),
+                io.LatentOperation.Input("operation"),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def patch(self, model, operation):
+    @classmethod
+    def execute(cls, model, operation) -> io.NodeOutput:
         m = model.clone()
 
         def pre_cfg_function(args):
@@ -261,21 +315,25 @@ class LatentApplyOperationCFG:
             return conds_out
 
         m.set_model_sampler_pre_cfg_function(pre_cfg_function)
-        return (m, )
+        return io.NodeOutput(m)
 
-class LatentOperationTonemapReinhard:
+class LatentOperationTonemapReinhard(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "multiplier": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 100.0, "step": 0.01}),
-                              }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentOperationTonemapReinhard",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Float.Input("multiplier", default=1.0, min=0.0, max=100.0, step=0.01),
+            ],
+            outputs=[
+                io.LatentOperation.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT_OPERATION",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def op(self, multiplier):
+    @classmethod
+    def execute(cls, multiplier) -> io.NodeOutput:
         def tonemap_reinhard(latent, **kwargs):
             latent_vector_magnitude = (torch.linalg.vector_norm(latent, dim=(1)) + 0.0000000001)[:,None]
             normalized_latent = latent / latent_vector_magnitude
@@ -291,39 +349,27 @@ class LatentOperationTonemapReinhard:
             new_magnitude *= top
 
             return normalized_latent * new_magnitude
-        return (tonemap_reinhard,)
+        return io.NodeOutput(tonemap_reinhard)
 
-class LatentOperationSharpen:
+class LatentOperationSharpen(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-                "sharpen_radius": ("INT", {
-                    "default": 9,
-                    "min": 1,
-                    "max": 31,
-                    "step": 1
-                }),
-                "sigma": ("FLOAT", {
-                    "default": 1.0,
-                    "min": 0.1,
-                    "max": 10.0,
-                    "step": 0.1
-                }),
-                "alpha": ("FLOAT", {
-                    "default": 0.1,
-                    "min": 0.0,
-                    "max": 5.0,
-                    "step": 0.01
-                }),
-                              }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentOperationSharpen",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Int.Input("sharpen_radius", default=9, min=1, max=31, step=1),
+                io.Float.Input("sigma", default=1.0, min=0.1, max=10.0, step=0.1),
+                io.Float.Input("alpha", default=0.1, min=0.0, max=5.0, step=0.01),
+            ],
+            outputs=[
+                io.LatentOperation.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT_OPERATION",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def op(self, sharpen_radius, sigma, alpha):
+    @classmethod
+    def execute(cls, sharpen_radius, sigma, alpha) -> io.NodeOutput:
         def sharpen(latent, **kwargs):
             luminance = (torch.linalg.vector_norm(latent, dim=(1)) + 1e-6)[:,None]
             normalized_latent = latent / luminance
@@ -340,19 +386,27 @@ class LatentOperationSharpen:
             sharpened = torch.nn.functional.conv2d(padded_image, kernel.repeat(channels, 1, 1).unsqueeze(1), padding=kernel_size // 2, groups=channels)[:,:,sharpen_radius:-sharpen_radius, sharpen_radius:-sharpen_radius]
 
             return luminance * sharpened
-        return (sharpen,)
+        return io.NodeOutput(sharpen)
 
-NODE_CLASS_MAPPINGS = {
-    "LatentAdd": LatentAdd,
-    "LatentSubtract": LatentSubtract,
-    "LatentMultiply": LatentMultiply,
-    "LatentInterpolate": LatentInterpolate,
-    "LatentConcat": LatentConcat,
-    "LatentCut": LatentCut,
-    "LatentBatch": LatentBatch,
-    "LatentBatchSeedBehavior": LatentBatchSeedBehavior,
-    "LatentApplyOperation": LatentApplyOperation,
-    "LatentApplyOperationCFG": LatentApplyOperationCFG,
-    "LatentOperationTonemapReinhard": LatentOperationTonemapReinhard,
-    "LatentOperationSharpen": LatentOperationSharpen,
-}
+
+class LatentExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            LatentAdd,
+            LatentSubtract,
+            LatentMultiply,
+            LatentInterpolate,
+            LatentConcat,
+            LatentCut,
+            LatentBatch,
+            LatentBatchSeedBehavior,
+            LatentApplyOperation,
+            LatentApplyOperationCFG,
+            LatentOperationTonemapReinhard,
+            LatentOperationSharpen,
+        ]
+
+
+async def comfy_entrypoint() -> LatentExtension:
+    return LatentExtension()

comfy_extras/nodes_lora_extract.py

@@ -5,6 +5,8 @@ import folder_paths
 import os
 import logging
 from enum import Enum
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 CLAMP_QUANTILE = 0.99
 
@@ -71,32 +73,40 @@ def calc_lora_model(model_diff, rank, prefix_model, prefix_lora, output_sd, lora
             output_sd["{}{}.diff_b".format(prefix_lora, k[len(prefix_model):-5])] = sd[k].contiguous().half().cpu()
     return output_sd
 
-class LoraSave:
-    def __init__(self):
-        self.output_dir = folder_paths.get_output_directory()
+class LoraSave(io.ComfyNode):
+    @classmethod
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LoraSave",
+            display_name="Extract and Save Lora",
+            category="_for_testing",
+            inputs=[
+                io.String.Input("filename_prefix", default="loras/ComfyUI_extracted_lora"),
+                io.Int.Input("rank", default=8, min=1, max=4096, step=1),
+                io.Combo.Input("lora_type", options=tuple(LORA_TYPES.keys())),
+                io.Boolean.Input("bias_diff", default=True),
+                io.Model.Input(
+                    "model_diff",
+                    tooltip="The ModelSubtract output to be converted to a lora.",
+                    optional=True,
+                ),
+                io.Clip.Input(
+                  "text_encoder_diff",
+                    tooltip="The CLIPSubtract output to be converted to a lora.",
+                    optional=True,
+                ),
+            ],
+            is_experimental=True,
+            is_output_node=True,
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"filename_prefix": ("STRING", {"default": "loras/ComfyUI_extracted_lora"}),
-                              "rank": ("INT", {"default": 8, "min": 1, "max": 4096, "step": 1}),
-                              "lora_type": (tuple(LORA_TYPES.keys()),),
-                              "bias_diff": ("BOOLEAN", {"default": True}),
-                            },
-                "optional": {"model_diff": ("MODEL", {"tooltip": "The ModelSubtract output to be converted to a lora."}),
-                             "text_encoder_diff": ("CLIP", {"tooltip": "The CLIPSubtract output to be converted to a lora."})},
-    }
-    RETURN_TYPES = ()
-    FUNCTION = "save"
-    OUTPUT_NODE = True
-
-    CATEGORY = "_for_testing"
-
-    def save(self, filename_prefix, rank, lora_type, bias_diff, model_diff=None, text_encoder_diff=None):
+    def execute(cls, filename_prefix, rank, lora_type, bias_diff, model_diff=None, text_encoder_diff=None) -> io.NodeOutput:
         if model_diff is None and text_encoder_diff is None:
-            return {}
+            return io.NodeOutput()
 
         lora_type = LORA_TYPES.get(lora_type)
-        full_output_folder, filename, counter, subfolder, filename_prefix = folder_paths.get_save_image_path(filename_prefix, self.output_dir)
+        full_output_folder, filename, counter, subfolder, filename_prefix = folder_paths.get_save_image_path(filename_prefix, folder_paths.get_output_directory())
 
         output_sd = {}
         if model_diff is not None:
@@ -108,12 +118,16 @@ class LoraSave:
         output_checkpoint = os.path.join(full_output_folder, output_checkpoint)
 
         comfy.utils.save_torch_file(output_sd, output_checkpoint, metadata=None)
-        return {}
+        return io.NodeOutput()
 
-NODE_CLASS_MAPPINGS = {
-    "LoraSave": LoraSave
-}
 
-NODE_DISPLAY_NAME_MAPPINGS = {
-    "LoraSave": "Extract and Save Lora"
-}
+class LoraSaveExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            LoraSave,
+        ]
+
+
+async def comfy_entrypoint() -> LoraSaveExtension:
+    return LoraSaveExtension()

comfy_extras/nodes_lt.py

@@ -34,6 +34,7 @@ class EmptyLTXVLatentVideo(io.ComfyNode):
         latent = torch.zeros([batch_size, 128, ((length - 1) // 8) + 1, height // 32, width // 32], device=comfy.model_management.intermediate_device())
         return io.NodeOutput({"samples": latent})
 
+    generate = execute  # TODO: remove
 
 class LTXVImgToVideo(io.ComfyNode):
     @classmethod
@@ -77,6 +78,8 @@ class LTXVImgToVideo(io.ComfyNode):
 
         return io.NodeOutput(positive, negative, {"samples": latent, "noise_mask": conditioning_latent_frames_mask})
 
+    generate = execute  # TODO: remove
+
 
 def conditioning_get_any_value(conditioning, key, default=None):
     for t in conditioning:
@@ -264,6 +267,8 @@ class LTXVAddGuide(io.ComfyNode):
 
         return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask})
 
+    generate = execute  # TODO: remove
+
 
 class LTXVCropGuides(io.ComfyNode):
     @classmethod
@@ -300,6 +305,8 @@ class LTXVCropGuides(io.ComfyNode):
 
         return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask})
 
+    crop = execute  # TODO: remove
+
 
 class LTXVConditioning(io.ComfyNode):
     @classmethod
@@ -498,6 +505,7 @@ class LTXVPreprocess(io.ComfyNode):
             output_images.append(preprocess(image[i], img_compression))
         return io.NodeOutput(torch.stack(output_images))
 
+    preprocess = execute  # TODO: remove
 
 class LtxvExtension(ComfyExtension):
     @override

comfy_extras/nodes_model_downscale.py

@@ -1,24 +1,33 @@
+from typing_extensions import override
 import comfy.utils
+from comfy_api.latest import ComfyExtension, io
 
-class PatchModelAddDownscale:
-    upscale_methods = ["bicubic", "nearest-exact", "bilinear", "area", "bislerp"]
+
+class PatchModelAddDownscale(io.ComfyNode):
+    UPSCALE_METHODS = ["bicubic", "nearest-exact", "bilinear", "area", "bislerp"]
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model": ("MODEL",),
-                              "block_number": ("INT", {"default": 3, "min": 1, "max": 32, "step": 1}),
-                              "downscale_factor": ("FLOAT", {"default": 2.0, "min": 0.1, "max": 9.0, "step": 0.001}),
-                              "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                              "end_percent": ("FLOAT", {"default": 0.35, "min": 0.0, "max": 1.0, "step": 0.001}),
-                              "downscale_after_skip": ("BOOLEAN", {"default": True}),
-                              "downscale_method": (s.upscale_methods,),
-                              "upscale_method": (s.upscale_methods,),
-                              }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "patch"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="PatchModelAddDownscale",
+            display_name="PatchModelAddDownscale (Kohya Deep Shrink)",
+            category="model_patches/unet",
+            inputs=[
+                io.Model.Input("model"),
+                io.Int.Input("block_number", default=3, min=1, max=32, step=1),
+                io.Float.Input("downscale_factor", default=2.0, min=0.1, max=9.0, step=0.001),
+                io.Float.Input("start_percent", default=0.0, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=0.35, min=0.0, max=1.0, step=0.001),
+                io.Boolean.Input("downscale_after_skip", default=True),
+                io.Combo.Input("downscale_method", options=cls.UPSCALE_METHODS),
+                io.Combo.Input("upscale_method", options=cls.UPSCALE_METHODS),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    CATEGORY = "model_patches/unet"
-
-    def patch(self, model, block_number, downscale_factor, start_percent, end_percent, downscale_after_skip, downscale_method, upscale_method):
+    @classmethod
+    def execute(cls, model, block_number, downscale_factor, start_percent, end_percent, downscale_after_skip, downscale_method, upscale_method) -> io.NodeOutput:
         model_sampling = model.get_model_object("model_sampling")
         sigma_start = model_sampling.percent_to_sigma(start_percent)
         sigma_end = model_sampling.percent_to_sigma(end_percent)
@@ -41,13 +50,21 @@ class PatchModelAddDownscale:
         else:
             m.set_model_input_block_patch(input_block_patch)
         m.set_model_output_block_patch(output_block_patch)
-        return (m, )
+        return io.NodeOutput(m)
 
-NODE_CLASS_MAPPINGS = {
-    "PatchModelAddDownscale": PatchModelAddDownscale,
-}
 
 NODE_DISPLAY_NAME_MAPPINGS = {
     # Sampling
-    "PatchModelAddDownscale": "PatchModelAddDownscale (Kohya Deep Shrink)",
+    "PatchModelAddDownscale": "",
 }
+
+class ModelDownscaleExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            PatchModelAddDownscale,
+        ]
+
+
+async def comfy_entrypoint() -> ModelDownscaleExtension:
+    return ModelDownscaleExtension()

comfy_extras/nodes_sd3.py

@@ -3,64 +3,83 @@ import comfy.sd
 import comfy.model_management
 import nodes
 import torch
-import comfy_extras.nodes_slg
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
+from comfy_extras.nodes_slg import SkipLayerGuidanceDiT
 
 
-class TripleCLIPLoader:
+class TripleCLIPLoader(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "clip_name1": (folder_paths.get_filename_list("text_encoders"), ), "clip_name2": (folder_paths.get_filename_list("text_encoders"), ), "clip_name3": (folder_paths.get_filename_list("text_encoders"), )
-                             }}
-    RETURN_TYPES = ("CLIP",)
-    FUNCTION = "load_clip"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="TripleCLIPLoader",
+            category="advanced/loaders",
+            description="[Recipes]\n\nsd3: clip-l, clip-g, t5",
+            inputs=[
+                io.Combo.Input("clip_name1", options=folder_paths.get_filename_list("text_encoders")),
+                io.Combo.Input("clip_name2", options=folder_paths.get_filename_list("text_encoders")),
+                io.Combo.Input("clip_name3", options=folder_paths.get_filename_list("text_encoders")),
+            ],
+            outputs=[
+                io.Clip.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/loaders"
-
-    DESCRIPTION = "[Recipes]\n\nsd3: clip-l, clip-g, t5"
-
-    def load_clip(self, clip_name1, clip_name2, clip_name3):
+    @classmethod
+    def execute(cls, clip_name1, clip_name2, clip_name3) -> io.NodeOutput:
         clip_path1 = folder_paths.get_full_path_or_raise("text_encoders", clip_name1)
         clip_path2 = folder_paths.get_full_path_or_raise("text_encoders", clip_name2)
         clip_path3 = folder_paths.get_full_path_or_raise("text_encoders", clip_name3)
         clip = comfy.sd.load_clip(ckpt_paths=[clip_path1, clip_path2, clip_path3], embedding_directory=folder_paths.get_folder_paths("embeddings"))
-        return (clip,)
+        return io.NodeOutput(clip)
+
+    load_clip = execute  # TODO: remove
 
 
-class EmptySD3LatentImage:
-    def __init__(self):
-        self.device = comfy.model_management.intermediate_device()
+class EmptySD3LatentImage(io.ComfyNode):
+    @classmethod
+    def define_schema(cls):
+        return io.Schema(
+            node_id="EmptySD3LatentImage",
+            category="latent/sd3",
+            inputs=[
+                io.Int.Input("width", default=1024, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("height", default=1024, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("batch_size", default=1, min=1, max=4096),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "width": ("INT", {"default": 1024, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                              "height": ("INT", {"default": 1024, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096})}}
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "generate"
+    def execute(cls, width, height, batch_size=1) -> io.NodeOutput:
+        latent = torch.zeros([batch_size, 16, height // 8, width // 8], device=comfy.model_management.intermediate_device())
+        return io.NodeOutput({"samples":latent})
 
-    CATEGORY = "latent/sd3"
-
-    def generate(self, width, height, batch_size=1):
-        latent = torch.zeros([batch_size, 16, height // 8, width // 8], device=self.device)
-        return ({"samples":latent}, )
+    generate = execute  # TODO: remove
 
 
-class CLIPTextEncodeSD3:
+class CLIPTextEncodeSD3(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "clip": ("CLIP", ),
-            "clip_l": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "clip_g": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "t5xxl": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "empty_padding": (["none", "empty_prompt"], )
-            }}
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "encode"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="CLIPTextEncodeSD3",
+            category="advanced/conditioning",
+            inputs=[
+                io.Clip.Input("clip"),
+                io.String.Input("clip_l", multiline=True, dynamic_prompts=True),
+                io.String.Input("clip_g", multiline=True, dynamic_prompts=True),
+                io.String.Input("t5xxl", multiline=True, dynamic_prompts=True),
+                io.Combo.Input("empty_padding", options=["none", "empty_prompt"]),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/conditioning"
-
-    def encode(self, clip, clip_l, clip_g, t5xxl, empty_padding):
+    @classmethod
+    def execute(cls, clip, clip_l, clip_g, t5xxl, empty_padding) -> io.NodeOutput:
         no_padding = empty_padding == "none"
 
         tokens = clip.tokenize(clip_g)
@@ -82,57 +101,112 @@ class CLIPTextEncodeSD3:
                 tokens["l"] += empty["l"]
             while len(tokens["l"]) > len(tokens["g"]):
                 tokens["g"] += empty["g"]
-        return (clip.encode_from_tokens_scheduled(tokens), )
+        return io.NodeOutput(clip.encode_from_tokens_scheduled(tokens))
+
+    encode = execute  # TODO: remove
 
 
-class ControlNetApplySD3(nodes.ControlNetApplyAdvanced):
+class ControlNetApplySD3(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"positive": ("CONDITIONING", ),
-                             "negative": ("CONDITIONING", ),
-                             "control_net": ("CONTROL_NET", ),
-                             "vae": ("VAE", ),
-                             "image": ("IMAGE", ),
-                             "strength": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 10.0, "step": 0.01}),
-                             "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.001})
-                             }}
-    CATEGORY = "conditioning/controlnet"
-    DEPRECATED = True
+    def define_schema(cls) -> io.Schema:
+        return io.Schema(
+            node_id="ControlNetApplySD3",
+            display_name="Apply Controlnet with VAE",
+            category="conditioning/controlnet",
+            inputs=[
+                io.Conditioning.Input("positive"),
+                io.Conditioning.Input("negative"),
+                io.ControlNet.Input("control_net"),
+                io.Vae.Input("vae"),
+                io.Image.Input("image"),
+                io.Float.Input("strength", default=1.0, min=0.0, max=10.0, step=0.01),
+                io.Float.Input("start_percent", default=0.0, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=1.0, min=0.0, max=1.0, step=0.001),
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Conditioning.Output(display_name="negative"),
+            ],
+            is_deprecated=True,
+        )
+
+    @classmethod
+    def execute(cls, positive, negative, control_net, image, strength, start_percent, end_percent, vae=None) -> io.NodeOutput:
+        if strength == 0:
+            return io.NodeOutput(positive, negative)
+
+        control_hint = image.movedim(-1, 1)
+        cnets = {}
+
+        out = []
+        for conditioning in [positive, negative]:
+            c = []
+            for t in conditioning:
+                d = t[1].copy()
+
+                prev_cnet = d.get('control', None)
+                if prev_cnet in cnets:
+                    c_net = cnets[prev_cnet]
+                else:
+                    c_net = control_net.copy().set_cond_hint(control_hint, strength, (start_percent, end_percent),
+                                                             vae=vae, extra_concat=[])
+                    c_net.set_previous_controlnet(prev_cnet)
+                    cnets[prev_cnet] = c_net
+
+                d['control'] = c_net
+                d['control_apply_to_uncond'] = False
+                n = [t[0], d]
+                c.append(n)
+            out.append(c)
+        return io.NodeOutput(out[0], out[1])
+
+    apply_controlnet = execute  # TODO: remove
 
 
-class SkipLayerGuidanceSD3(comfy_extras.nodes_slg.SkipLayerGuidanceDiT):
+class SkipLayerGuidanceSD3(io.ComfyNode):
     '''
     Enhance guidance towards detailed dtructure by having another set of CFG negative with skipped layers.
     Inspired by Perturbed Attention Guidance (https://arxiv.org/abs/2403.17377)
     Experimental implementation by Dango233@StabilityAI.
     '''
+
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"model": ("MODEL", ),
-                             "layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "scale": ("FLOAT", {"default": 3.0, "min": 0.0, "max": 10.0, "step": 0.1}),
-                             "start_percent": ("FLOAT", {"default": 0.01, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 0.15, "min": 0.0, "max": 1.0, "step": 0.001})
-                                }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "skip_guidance_sd3"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SkipLayerGuidanceSD3",
+            category="advanced/guidance",
+            description="Generic version of SkipLayerGuidance node that can be used on every DiT model.",
+            inputs=[
+                io.Model.Input("model"),
+                io.String.Input("layers", default="7, 8, 9", multiline=False),
+                io.Float.Input("scale", default=3.0, min=0.0, max=10.0, step=0.1),
+                io.Float.Input("start_percent", default=0.01, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=0.15, min=0.0, max=1.0, step=0.001),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+            is_experimental=True,
+        )
 
-    CATEGORY = "advanced/guidance"
+    @classmethod
+    def execute(cls, model, layers, scale, start_percent, end_percent) -> io.NodeOutput:
+        return SkipLayerGuidanceDiT().execute(model=model, scale=scale, start_percent=start_percent, end_percent=end_percent, double_layers=layers)
 
-    def skip_guidance_sd3(self, model, layers, scale, start_percent, end_percent):
-        return self.skip_guidance(model=model, scale=scale, start_percent=start_percent, end_percent=end_percent, double_layers=layers)
+    skip_guidance_sd3 = execute  # TODO: remove
 
 
-NODE_CLASS_MAPPINGS = {
-    "TripleCLIPLoader": TripleCLIPLoader,
-    "EmptySD3LatentImage": EmptySD3LatentImage,
-    "CLIPTextEncodeSD3": CLIPTextEncodeSD3,
-    "ControlNetApplySD3": ControlNetApplySD3,
-    "SkipLayerGuidanceSD3": SkipLayerGuidanceSD3,
-}
+class SD3Extension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            TripleCLIPLoader,
+            EmptySD3LatentImage,
+            CLIPTextEncodeSD3,
+            ControlNetApplySD3,
+            SkipLayerGuidanceSD3,
+        ]
 
-NODE_DISPLAY_NAME_MAPPINGS = {
-    # Sampling
-    "ControlNetApplySD3": "Apply Controlnet with VAE",
-}
+
+async def comfy_entrypoint() -> SD3Extension:
+    return SD3Extension()

comfy_extras/nodes_slg.py

@@ -1,33 +1,40 @@
 import comfy.model_patcher
 import comfy.samplers
 import re
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 
-class SkipLayerGuidanceDiT:
+class SkipLayerGuidanceDiT(io.ComfyNode):
     '''
     Enhance guidance towards detailed dtructure by having another set of CFG negative with skipped layers.
     Inspired by Perturbed Attention Guidance (https://arxiv.org/abs/2403.17377)
     Original experimental implementation for SD3 by Dango233@StabilityAI.
     '''
+
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"model": ("MODEL", ),
-                             "double_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "single_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "scale": ("FLOAT", {"default": 3.0, "min": 0.0, "max": 10.0, "step": 0.1}),
-                             "start_percent": ("FLOAT", {"default": 0.01, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 0.15, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "rescaling_scale": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 10.0, "step": 0.01}),
-                                }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "skip_guidance"
-    EXPERIMENTAL = True
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SkipLayerGuidanceDiT",
+            category="advanced/guidance",
+            description="Generic version of SkipLayerGuidance node that can be used on every DiT model.",
+            is_experimental=True,
+            inputs=[
+                io.Model.Input("model"),
+                io.String.Input("double_layers", default="7, 8, 9"),
+                io.String.Input("single_layers", default="7, 8, 9"),
+                io.Float.Input("scale", default=3.0, min=0.0, max=10.0, step=0.1),
+                io.Float.Input("start_percent", default=0.01, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=0.15, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("rescaling_scale", default=0.0, min=0.0, max=10.0, step=0.01),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    DESCRIPTION = "Generic version of SkipLayerGuidance node that can be used on every DiT model."
-
-    CATEGORY = "advanced/guidance"
-
-    def skip_guidance(self, model, scale, start_percent, end_percent, double_layers="", single_layers="", rescaling_scale=0):
+    @classmethod
+    def execute(cls, model, scale, start_percent, end_percent, double_layers="", single_layers="", rescaling_scale=0) -> io.NodeOutput:
         # check if layer is comma separated integers
         def skip(args, extra_args):
             return args
@@ -43,7 +50,7 @@ class SkipLayerGuidanceDiT:
         single_layers = [int(i) for i in single_layers]
 
         if len(double_layers) == 0 and len(single_layers) == 0:
-            return (model, )
+            return io.NodeOutput(model)
 
         def post_cfg_function(args):
             model = args["model"]
@@ -76,29 +83,36 @@ class SkipLayerGuidanceDiT:
         m = model.clone()
         m.set_model_sampler_post_cfg_function(post_cfg_function)
 
-        return (m, )
+        return io.NodeOutput(m)
 
-class SkipLayerGuidanceDiTSimple:
+    skip_guidance = execute  # TODO: remove
+
+
+class SkipLayerGuidanceDiTSimple(io.ComfyNode):
     '''
     Simple version of the SkipLayerGuidanceDiT node that only modifies the uncond pass.
     '''
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"model": ("MODEL", ),
-                             "double_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "single_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                                }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "skip_guidance"
-    EXPERIMENTAL = True
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SkipLayerGuidanceDiTSimple",
+            category="advanced/guidance",
+            description="Simple version of the SkipLayerGuidanceDiT node that only modifies the uncond pass.",
+            is_experimental=True,
+            inputs=[
+                io.Model.Input("model"),
+                io.String.Input("double_layers", default="7, 8, 9"),
+                io.String.Input("single_layers", default="7, 8, 9"),
+                io.Float.Input("start_percent", default=0.0, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=1.0, min=0.0, max=1.0, step=0.001),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    DESCRIPTION = "Simple version of the SkipLayerGuidanceDiT node that only modifies the uncond pass."
-
-    CATEGORY = "advanced/guidance"
-
-    def skip_guidance(self, model, start_percent, end_percent, double_layers="", single_layers=""):
+    @classmethod
+    def execute(cls, model, start_percent, end_percent, double_layers="", single_layers="") -> io.NodeOutput:
         def skip(args, extra_args):
             return args
 
@@ -113,7 +127,7 @@ class SkipLayerGuidanceDiTSimple:
         single_layers = [int(i) for i in single_layers]
 
         if len(double_layers) == 0 and len(single_layers) == 0:
-            return (model, )
+            return io.NodeOutput(model)
 
         def calc_cond_batch_function(args):
             x = args["input"]
@@ -144,9 +158,19 @@ class SkipLayerGuidanceDiTSimple:
         m = model.clone()
         m.set_model_sampler_calc_cond_batch_function(calc_cond_batch_function)
 
-        return (m, )
+        return io.NodeOutput(m)
 
-NODE_CLASS_MAPPINGS = {
-    "SkipLayerGuidanceDiT": SkipLayerGuidanceDiT,
-    "SkipLayerGuidanceDiTSimple": SkipLayerGuidanceDiTSimple,
-}
+    skip_guidance = execute  # TODO: remove
+
+
+class SkipLayerGuidanceExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            SkipLayerGuidanceDiT,
+            SkipLayerGuidanceDiTSimple,
+        ]
+
+
+async def comfy_entrypoint() -> SkipLayerGuidanceExtension:
+    return SkipLayerGuidanceExtension()

comfy_extras/nodes_stable3d.py

@@ -1,6 +1,8 @@
 import torch
 import nodes
 import comfy.utils
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 def camera_embeddings(elevation, azimuth):
     elevation = torch.as_tensor([elevation])
@@ -20,26 +22,31 @@ def camera_embeddings(elevation, azimuth):
     return embeddings
 
 
-class StableZero123_Conditioning:
+class StableZero123_Conditioning(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "clip_vision": ("CLIP_VISION",),
-                              "init_image": ("IMAGE",),
-                              "vae": ("VAE",),
-                              "width": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}),
-                              "height": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}),
-                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}),
-                              "elevation": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}),
-                              "azimuth": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}),
-                             }}
-    RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT")
-    RETURN_NAMES = ("positive", "negative", "latent")
+    def define_schema(cls):
+        return io.Schema(
+            node_id="StableZero123_Conditioning",
+            category="conditioning/3d_models",
+            inputs=[
+                io.ClipVision.Input("clip_vision"),
+                io.Image.Input("init_image"),
+                io.Vae.Input("vae"),
+                io.Int.Input("width", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8),
+                io.Int.Input("height", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8),
+                io.Int.Input("batch_size", default=1, min=1, max=4096),
+                io.Float.Input("elevation", default=0.0, min=-180.0, max=180.0, step=0.1, round=False),
+                io.Float.Input("azimuth", default=0.0, min=-180.0, max=180.0, step=0.1, round=False)
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Conditioning.Output(display_name="negative"),
+                io.Latent.Output(display_name="latent")
+            ]
+        )
 
-    FUNCTION = "encode"
-
-    CATEGORY = "conditioning/3d_models"
-
-    def encode(self, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth):
+    @classmethod
+    def execute(cls, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth) -> io.NodeOutput:
         output = clip_vision.encode_image(init_image)
         pooled = output.image_embeds.unsqueeze(0)
         pixels = comfy.utils.common_upscale(init_image.movedim(-1,1), width, height, "bilinear", "center").movedim(1,-1)
@@ -51,30 +58,35 @@ class StableZero123_Conditioning:
         positive = [[cond, {"concat_latent_image": t}]]
         negative = [[torch.zeros_like(pooled), {"concat_latent_image": torch.zeros_like(t)}]]
         latent = torch.zeros([batch_size, 4, height // 8, width // 8])
-        return (positive, negative, {"samples":latent})
+        return io.NodeOutput(positive, negative, {"samples":latent})
 
-class StableZero123_Conditioning_Batched:
+class StableZero123_Conditioning_Batched(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "clip_vision": ("CLIP_VISION",),
-                              "init_image": ("IMAGE",),
-                              "vae": ("VAE",),
-                              "width": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}),
-                              "height": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}),
-                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}),
-                              "elevation": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}),
-                              "azimuth": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}),
-                              "elevation_batch_increment": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}),
-                              "azimuth_batch_increment": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}),
-                             }}
-    RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT")
-    RETURN_NAMES = ("positive", "negative", "latent")
+    def define_schema(cls):
+        return io.Schema(
+            node_id="StableZero123_Conditioning_Batched",
+            category="conditioning/3d_models",
+            inputs=[
+                io.ClipVision.Input("clip_vision"),
+                io.Image.Input("init_image"),
+                io.Vae.Input("vae"),
+                io.Int.Input("width", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8),
+                io.Int.Input("height", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8),
+                io.Int.Input("batch_size", default=1, min=1, max=4096),
+                io.Float.Input("elevation", default=0.0, min=-180.0, max=180.0, step=0.1, round=False),
+                io.Float.Input("azimuth", default=0.0, min=-180.0, max=180.0, step=0.1, round=False),
+                io.Float.Input("elevation_batch_increment", default=0.0, min=-180.0, max=180.0, step=0.1, round=False),
+                io.Float.Input("azimuth_batch_increment", default=0.0, min=-180.0, max=180.0, step=0.1, round=False)
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Conditioning.Output(display_name="negative"),
+                io.Latent.Output(display_name="latent")
+            ]
+        )
 
-    FUNCTION = "encode"
-
-    CATEGORY = "conditioning/3d_models"
-
-    def encode(self, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth, elevation_batch_increment, azimuth_batch_increment):
+    @classmethod
+    def execute(cls, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth, elevation_batch_increment, azimuth_batch_increment) -> io.NodeOutput:
         output = clip_vision.encode_image(init_image)
         pooled = output.image_embeds.unsqueeze(0)
         pixels = comfy.utils.common_upscale(init_image.movedim(-1,1), width, height, "bilinear", "center").movedim(1,-1)
@@ -93,27 +105,32 @@ class StableZero123_Conditioning_Batched:
         positive = [[cond, {"concat_latent_image": t}]]
         negative = [[torch.zeros_like(pooled), {"concat_latent_image": torch.zeros_like(t)}]]
         latent = torch.zeros([batch_size, 4, height // 8, width // 8])
-        return (positive, negative, {"samples":latent, "batch_index": [0] * batch_size})
+        return io.NodeOutput(positive, negative, {"samples":latent, "batch_index": [0] * batch_size})
 
-class SV3D_Conditioning:
+class SV3D_Conditioning(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "clip_vision": ("CLIP_VISION",),
-                              "init_image": ("IMAGE",),
-                              "vae": ("VAE",),
-                              "width": ("INT", {"default": 576, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}),
-                              "height": ("INT", {"default": 576, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}),
-                              "video_frames": ("INT", {"default": 21, "min": 1, "max": 4096}),
-                              "elevation": ("FLOAT", {"default": 0.0, "min": -90.0, "max": 90.0, "step": 0.1, "round": False}),
-                             }}
-    RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT")
-    RETURN_NAMES = ("positive", "negative", "latent")
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SV3D_Conditioning",
+            category="conditioning/3d_models",
+            inputs=[
+                io.ClipVision.Input("clip_vision"),
+                io.Image.Input("init_image"),
+                io.Vae.Input("vae"),
+                io.Int.Input("width", default=576, min=16, max=nodes.MAX_RESOLUTION, step=8),
+                io.Int.Input("height", default=576, min=16, max=nodes.MAX_RESOLUTION, step=8),
+                io.Int.Input("video_frames", default=21, min=1, max=4096),
+                io.Float.Input("elevation", default=0.0, min=-90.0, max=90.0, step=0.1, round=False)
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Conditioning.Output(display_name="negative"),
+                io.Latent.Output(display_name="latent")
+            ]
+        )
 
-    FUNCTION = "encode"
-
-    CATEGORY = "conditioning/3d_models"
-
-    def encode(self, clip_vision, init_image, vae, width, height, video_frames, elevation):
+    @classmethod
+    def execute(cls, clip_vision, init_image, vae, width, height, video_frames, elevation) -> io.NodeOutput:
         output = clip_vision.encode_image(init_image)
         pooled = output.image_embeds.unsqueeze(0)
         pixels = comfy.utils.common_upscale(init_image.movedim(-1,1), width, height, "bilinear", "center").movedim(1,-1)
@@ -133,11 +150,17 @@ class SV3D_Conditioning:
         positive = [[pooled, {"concat_latent_image": t, "elevation": elevations, "azimuth": azimuths}]]
         negative = [[torch.zeros_like(pooled), {"concat_latent_image": torch.zeros_like(t), "elevation": elevations, "azimuth": azimuths}]]
         latent = torch.zeros([video_frames, 4, height // 8, width // 8])
-        return (positive, negative, {"samples":latent})
+        return io.NodeOutput(positive, negative, {"samples":latent})
 
 
-NODE_CLASS_MAPPINGS = {
-    "StableZero123_Conditioning": StableZero123_Conditioning,
-    "StableZero123_Conditioning_Batched": StableZero123_Conditioning_Batched,
-    "SV3D_Conditioning": SV3D_Conditioning,
-}
+class Stable3DExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            StableZero123_Conditioning,
+            StableZero123_Conditioning_Batched,
+            SV3D_Conditioning,
+        ]
+
+async def comfy_entrypoint() -> Stable3DExtension:
+    return Stable3DExtension()

comfy_extras/nodes_tomesd.py

@@ -1,7 +1,9 @@
 #Taken from: https://github.com/dbolya/tomesd
 
 import torch
-from typing import Tuple, Callable
+from typing import Tuple, Callable, Optional
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 import math
 
 def do_nothing(x: torch.Tensor, mode:str=None):
@@ -144,33 +146,45 @@ def get_functions(x, ratio, original_shape):
 
 
 
-class TomePatchModel:
+class TomePatchModel(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model": ("MODEL",),
-                              "ratio": ("FLOAT", {"default": 0.3, "min": 0.0, "max": 1.0, "step": 0.01}),
-                              }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "patch"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="TomePatchModel",
+            category="model_patches/unet",
+            inputs=[
+                io.Model.Input("model"),
+                io.Float.Input("ratio", default=0.3, min=0.0, max=1.0, step=0.01),
+            ],
+            outputs=[io.Model.Output()],
+        )
 
-    CATEGORY = "model_patches/unet"
-
-    def patch(self, model, ratio):
-        self.u = None
+    @classmethod
+    def execute(cls, model, ratio) -> io.NodeOutput:
+        u: Optional[Callable] = None
         def tomesd_m(q, k, v, extra_options):
+            nonlocal u
             #NOTE: In the reference code get_functions takes x (input of the transformer block) as the argument instead of q
             #however from my basic testing it seems that using q instead gives better results
-            m, self.u = get_functions(q, ratio, extra_options["original_shape"])
+            m, u = get_functions(q, ratio, extra_options["original_shape"])
             return m(q), k, v
         def tomesd_u(n, extra_options):
-            return self.u(n)
+            nonlocal u
+            return u(n)
 
         m = model.clone()
         m.set_model_attn1_patch(tomesd_m)
         m.set_model_attn1_output_patch(tomesd_u)
-        return (m, )
+        return io.NodeOutput(m)
 
 
-NODE_CLASS_MAPPINGS = {
-    "TomePatchModel": TomePatchModel,
-}
+class TomePatchModelExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            TomePatchModel,
+        ]
+
+
+async def comfy_entrypoint() -> TomePatchModelExtension:
+    return TomePatchModelExtension()

comfy_extras/nodes_torch_compile.py

@@ -1,23 +1,39 @@
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 from comfy_api.torch_helpers import set_torch_compile_wrapper
 
 
-class TorchCompileModel:
+class TorchCompileModel(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model": ("MODEL",),
-                             "backend": (["inductor", "cudagraphs"],),
-                              }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "patch"
+    def define_schema(cls) -> io.Schema:
+        return io.Schema(
+            node_id="TorchCompileModel",
+            category="_for_testing",
+            inputs=[
+                io.Model.Input("model"),
+                io.Combo.Input(
+                    "backend",
+                    options=["inductor", "cudagraphs"],
+                ),
+            ],
+            outputs=[io.Model.Output()],
+            is_experimental=True,
+        )
 
-    CATEGORY = "_for_testing"
-    EXPERIMENTAL = True
-
-    def patch(self, model, backend):
+    @classmethod
+    def execute(cls, model, backend) -> io.NodeOutput:
         m = model.clone()
         set_torch_compile_wrapper(model=m, backend=backend)
-        return (m, )
+        return io.NodeOutput(m)
 
-NODE_CLASS_MAPPINGS = {
-    "TorchCompileModel": TorchCompileModel,
-}
+
+class TorchCompileExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            TorchCompileModel,
+        ]
+
+
+async def comfy_entrypoint() -> TorchCompileExtension:
+    return TorchCompileExtension()

comfy_extras/nodes_upscale_model.py

@@ -4,6 +4,8 @@ from comfy import model_management
 import torch
 import comfy.utils
 import folder_paths
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 try:
     from spandrel_extra_arches import EXTRA_REGISTRY
@@ -13,17 +15,23 @@ try:
 except:
     pass
 
-class UpscaleModelLoader:
+class UpscaleModelLoader(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model_name": (folder_paths.get_filename_list("upscale_models"), ),
-                             }}
-    RETURN_TYPES = ("UPSCALE_MODEL",)
-    FUNCTION = "load_model"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="UpscaleModelLoader",
+            display_name="Load Upscale Model",
+            category="loaders",
+            inputs=[
+                io.Combo.Input("model_name", options=folder_paths.get_filename_list("upscale_models")),
+            ],
+            outputs=[
+                io.UpscaleModel.Output(),
+            ],
+        )
 
-    CATEGORY = "loaders"
-
-    def load_model(self, model_name):
+    @classmethod
+    def execute(cls, model_name) -> io.NodeOutput:
         model_path = folder_paths.get_full_path_or_raise("upscale_models", model_name)
         sd = comfy.utils.load_torch_file(model_path, safe_load=True)
         if "module.layers.0.residual_group.blocks.0.norm1.weight" in sd:
@@ -33,21 +41,29 @@ class UpscaleModelLoader:
         if not isinstance(out, ImageModelDescriptor):
             raise Exception("Upscale model must be a single-image model.")
 
-        return (out, )
+        return io.NodeOutput(out)
+
+    load_model = execute  # TODO: remove
 
 
-class ImageUpscaleWithModel:
+class ImageUpscaleWithModel(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "upscale_model": ("UPSCALE_MODEL",),
-                              "image": ("IMAGE",),
-                              }}
-    RETURN_TYPES = ("IMAGE",)
-    FUNCTION = "upscale"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="ImageUpscaleWithModel",
+            display_name="Upscale Image (using Model)",
+            category="image/upscaling",
+            inputs=[
+                io.UpscaleModel.Input("upscale_model"),
+                io.Image.Input("image"),
+            ],
+            outputs=[
+                io.Image.Output(),
+            ],
+        )
 
-    CATEGORY = "image/upscaling"
-
-    def upscale(self, upscale_model, image):
+    @classmethod
+    def execute(cls, upscale_model, image) -> io.NodeOutput:
         device = model_management.get_torch_device()
 
         memory_required = model_management.module_size(upscale_model.model)
@@ -75,9 +91,19 @@ class ImageUpscaleWithModel:
 
         upscale_model.to("cpu")
         s = torch.clamp(s.movedim(-3,-1), min=0, max=1.0)
-        return (s,)
+        return io.NodeOutput(s)
 
-NODE_CLASS_MAPPINGS = {
-    "UpscaleModelLoader": UpscaleModelLoader,
-    "ImageUpscaleWithModel": ImageUpscaleWithModel
-}
+    upscale = execute  # TODO: remove
+
+
+class UpscaleModelExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            UpscaleModelLoader,
+            ImageUpscaleWithModel,
+        ]
+
+
+async def comfy_entrypoint() -> UpscaleModelExtension:
+    return UpscaleModelExtension()

comfyui_version.py

@@ -1,3 +1,3 @@
 # This file is automatically generated by the build process when version is
 # updated in pyproject.toml.
-__version__ = "0.3.62"
+__version__ = "0.3.64"

nodes.py

@@ -2027,7 +2027,6 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "DiffControlNetLoader": "Load ControlNet Model (diff)",
     "StyleModelLoader": "Load Style Model",
     "CLIPVisionLoader": "Load CLIP Vision",
-    "UpscaleModelLoader": "Load Upscale Model",
     "UNETLoader": "Load Diffusion Model",
     # Conditioning
     "CLIPVisionEncode": "CLIP Vision Encode",
@@ -2065,7 +2064,6 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "LoadImageOutput": "Load Image (from Outputs)",
     "ImageScale": "Upscale Image",
     "ImageScaleBy": "Upscale Image By",
-    "ImageUpscaleWithModel": "Upscale Image (using Model)",
     "ImageInvert": "Invert Image",
     "ImagePadForOutpaint": "Pad Image for Outpainting",
     "ImageBatch": "Batch Images",
@@ -2357,6 +2355,7 @@ async def init_builtin_api_nodes():
         "nodes_stability.py",
         "nodes_pika.py",
         "nodes_runway.py",
+        "nodes_sora.py",
         "nodes_tripo.py",
         "nodes_moonvalley.py",
         "nodes_rodin.py",

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "ComfyUI"
-version = "0.3.62"
+version = "0.3.64"
 readme = "README.md"
 license = { file = "LICENSE" }
 requires-python = ">=3.9"
@@ -57,22 +57,13 @@ messages_control.disable = [
   "redefined-builtin",
   "unnecessary-lambda",
   "dangerous-default-value",
+  "invalid-overridden-method",
   # next warnings should be fixed in future
   "bad-classmethod-argument",  # Class method should have 'cls' as first argument
   "wrong-import-order",  # Standard imports should be placed before third party imports
-  "logging-fstring-interpolation", # Use lazy % formatting in logging functions
   "ungrouped-imports",
   "unnecessary-pass",
-  "unidiomatic-typecheck",
   "unnecessary-lambda-assignment",
-  "bad-indentation",
   "no-else-return",
-  "no-else-raise",
-  "invalid-overridden-method",
   "unused-variable",
-  "pointless-string-statement",
-  "inconsistent-return-statements",
-  "import-outside-toplevel",
-  "reimported",
-  "redefined-outer-name",
 ]

requirements.txt

@@ -1,5 +1,5 @@
-comfyui-frontend-package==1.27.7
-comfyui-workflow-templates==0.1.91
+comfyui-frontend-package==1.27.10
+comfyui-workflow-templates==0.1.94
 comfyui-embedded-docs==0.2.6
 torch
 torchsde
@@ -25,6 +25,5 @@ av>=14.2.0
 #non essential dependencies:
 kornia>=0.7.1
 spandrel
-soundfile
 pydantic~=2.0
 pydantic-settings~=2.0