Comfy-Org/ComfyUI · error · ValueError

Could not determine frame count for file '{self.__file}'\nNo

Error message

Could not determine frame count for file '{self.__file}'\nNo frames exist for start_time {self.__start_time}

What it means

When counting frames in a trim window, VideoFromFile seeks to start_pts and iterates until the first frame with pts >= start_pts. If the iterator exhausts without any frame reaching start_pts — the requested start_time lies beyond the last frame of the stream — the for/else raises, combining 'Could not determine frame count' with the fact that no frames exist at that start_time.

Source

Thrown at comfy_api/latest/_input_impl/video_types.py:266

                if estimated_frames > 0:
                    return estimated_frames

            # 3. Last resort: decode frames and count them (streaming)
            start_time, duration = self.get_active_trim_window()
            frame_count = 1
            start_pts = int(start_time / video_stream.time_base)
            end_pts = int((start_time + duration) / video_stream.time_base)
            container.seek(start_pts, stream=video_stream)
            frame_iterator = (
                container.decode(video_stream)
                if video_stream.codec.capabilities & 0x100
                else container.demux(video_stream)
            )
            for frame in frame_iterator:
                if frame.pts >= start_pts:
                    break
            else:
                raise ValueError(f"Could not determine frame count for file '{self.__file}'\nNo frames exist for start_time {self.__start_time}")
            for frame in frame_iterator:
                if frame.pts >= end_pts:
                    break
                frame_count += 1
            return frame_count

    def get_frame_rate(self) -> Fraction:
        """
        Returns the average frame rate of the video using container metadata
        without decoding all frames.
        """
        if isinstance(self.__file, io.BytesIO):
            self.__file.seek(0)

        with av.open(self.__file, mode="r") as container:
            video_stream = self._get_first_video_stream(container)
            # Preferred: use PyAV's average_rate (usually already a Fraction-like)
            if video_stream.average_rate:

View on GitHub (pinned to 1c6d8d45b3)

Solutions

1. Clamp start_time below the video's actual duration: start = min(start_time, get_duration() - 1/fps)
2. Validate trim parameters against get_duration() before calling get_frame_count/as_trimmed
3. If metadata duration was wrong, remux to fix the header (see duration error) so clamping works off real values

Example fix

// before
count = video.get_frame_count(start_time=10.0)  # clip is 8s long

# after
dur = video.get_duration()
count = video.get_frame_count(start_time=min(10.0, dur - 0.1))

dur = video.get_duration()
if start_time >= dur:
    start_time = max(0.0, dur - 1.0 / float(video.get_frame_rate()))

When it happens

Trigger: Calling get_frame_count (or as_trimmed with a window) where start_time is at or after the end of the video, e.g. start_time >= duration, so every decoded frame has pts < start_pts.

Common situations: Trim node computes start_time from a larger assumed duration; user requests the last fraction of a clip whose real duration is shorter than metadata claims; off-by-one seek to end_pts exactly.

Related errors

• Custom alpha video frame count ({mask.shape[0]}) does not ma
• Video too short: need at least 16 frames for Moonvalley
• Failed to trim video: {str(e)}
• resize produced no frames (start_time={start_time}, duration
• {type(model).__name__} must implement map_context_window_to_

AI-assisted analysis of Comfy-Org/ComfyUI@1c6d8d45b3 (2026-08-14). Data as JSON: /api/errors/f02196b11e712d89. Report an issue: GitHub.