Comfy-Org/ComfyUI · warning · ValueError

Currently only one last frame image is supported.

Error message

Currently only one last frame image is supported.

What it means

Client-side ValueError in the v2.5 LTX image-to-video node: the optional last_frame input, when connected, must also contain exactly one image. Raised only when last_frame is not None and its image count != 1.

Source

Thrown at comfy_api_nodes/nodes_ltxv.py:484

        )

    @classmethod
    async def execute(
        cls,
        image: Input.Image,
        model: dict,
        prompt: str,
        seed: int = 42,
        last_frame: Input.Image | None = None,
    ) -> IO.NodeOutput:
        validate_string(prompt, min_length=1, max_length=10000)
        _v25_validate_settings(model)
        if get_number_of_images(image) != 1:
            raise ValueError("Currently only one input image is supported.")
        last_frame_uri = None
        if last_frame is not None:
            if get_number_of_images(last_frame) != 1:
                raise ValueError("Currently only one last frame image is supported.")
            last_frame_uri = (await upload_images_to_comfyapi(cls, last_frame, max_images=1, mime_type="image/png"))[0]
        return await _v25_submit_and_poll(
            cls,
            "image-to-video",
            ExecuteTaskRequest(
                image_uri=(await upload_images_to_comfyapi(cls, image, max_images=1, mime_type="image/png"))[0],
                last_frame_uri=last_frame_uri,
                prompt=prompt,
                model=V25_MODELS_MAP[model["model"]],
                duration=int(model["duration"]),
                resolution=model["resolution"],
                fps=int(model["fps"]),
                generate_audio=model["generate_audio"],
            ),
        )


class Ltx25AudioToVideoNode(IO.ComfyNode):

View on GitHub (pinned to 1c6d8d45b3)

Solutions

1. Reduce the last_frame input to a single image (batch 1 or pick one frame).
2. If last frame isn't needed, disconnect the input entirely — it is optional.
3. Use the same single-image discipline for both image and last_frame inputs.

if last_frame is not None:
    assert last_frame.shape[0] == 1, "last_frame must be a single image"

def is_single_image(t: torch.Tensor | None) -> bool:
    return t is None or t.shape[0] == 1

When it happens

Trigger: Connecting a batched or empty IMAGE tensor to the last_frame input of the v2.5 image-to-video node — e.g. both start and end frames loaded as batches, or last_frame sourced from a video frame stack.

Common situations: First/last-frame workflows where both endpoints come from batch loaders; last frame from a decoded video where the whole stack is connected instead of one frame.

Related errors

• Currently only one input image is supported.
• Durations over 10s require a 720p or 1080p resolution and 24
• Durations over 10s are only available for the Fast model at
• Audio duration must be between 2 and 20 seconds, got {audio_
• Total storyboard duration ({duration}s) must be between 3 an

AI-assisted analysis of Comfy-Org/ComfyUI@1c6d8d45b3 (2026-08-14). Data as JSON: /api/errors/385a26ab5b86a728. Report an issue: GitHub.