agentic-video Gallery

Quote what a tutorial narrator says at one moment

What was saidTutorialVerified
0:30CC-BY-3.0 · Living Colour Filmproduktion, Hamburg, for Wikimedia Deutschland

Proof frames

Click a frame to seek

Tool calls

3 calls · agentic-video 0.3.1
transcribeimac · 75 ms

Arguments

{
  "source": "$CLIP"
}

Result

{
  "ok": true,
  "data": {
    "status": "in_progress",
    "requested_model_size": "tiny",
    "requested_model_cached": false,
    "transcription_status": "started",
    "retry_after_seconds": 15,
    "transcription_message": "The tiny transcription started in the background. About 0% done (estimate): 0 s elapsed of roughly 10 s. Retry in about 15 seconds.",
    "elapsed_sec": 0,
    "estimated_total_sec": 10,
    "estimated_percent": 0
  }
}
transcribeimac · 2 ms

Arguments

{
  "source": "$CLIP"
}

Result

{
  "ok": true,
  "data": {
    "status": "in_progress",
    "requested_model_size": "tiny",
    "requested_model_cached": false,
    "transcription_status": "in_progress",
    "retry_after_seconds": 15,
    "transcription_message": "A transcription for this video is already running. About 30% done (estimate): 3 s elapsed of roughly 10 s. Retry in about 15 seconds.",
    "elapsed_sec": 3,
    "estimated_total_sec": 10,
    "estimated_percent": 30
  }
}
transcribeimac · 3 ms

Arguments

{
  "source": "$CLIP"
}

Result

{
  "ok": true,
  "data": {
    "source_path": "~/hill/gallery/repo/cards/quote-tutorial/clip.mp4",
    "language": "en",
    "duration_sec": 29.96,
    "segment_count": 6,
    "segments": [
      {
        "id": 1,
        "start_sec": 1.33,
        "end_sec": 3.83,
        "text": "You will see all of the changes that have been made.",
        "start_timecode": "00:00:01.330",
        "end_timecode": "00:00:03.830",
        "confidence": 0.954
      },
      {
        "id": 2,
        "start_sec": 3.83,
        "end_sec": 9.33,
        "text": "What you see here are the previous version and the new modified version side by side.",
        "start_timecode": "00:00:03.830",
        "end_timecode": "00:00:09.330",
        "confidence": 0.954
      },
      {
        "id": 3,
        "start_sec": 9.33,
        "end_sec": 13.33,
        "text": "This way, all of the changes are totally transparent.",
        "start_timecode": "00:00:09.330",
        "end_timecode": "00:00:13.330",
        "confidence": 0.954
      },
      {
        "id": 4,
        "start_sec": 13.33,
        "end_sec": 19.33,
        "text": "The alert and active community will carefully watch for modifications and correct any false information.",
        "start_timecode": "00:00:13.330",
        "end_timecode": "00:00:19.330",
        "confidence": 0.954
      },
      {
        "id": 5,
        "start_sec": 19.33,
        "end_sec": 25.33,
        "text": "This is just one of many effective ways to maintain the integrity of Wikipedia's content.",
        "start_timecode": "00:00:19.330",
        "end_timecode": "00:00:25.330",
        "confidence": 0.954
      },
      {
        "id": 6,
        "start_sec": 25.33,
        "end_sec": 30.33,
        "text": "The other tool I want to tell you about is a personalized recent changes list, or what...",
        "start_timecode": "00:00:25.330",
        "end_timecode": "00:00:30.330",
        "confidence": 0.954
      }
    ],
    "has_speech": true,
    "notice": "Note: Non-speech audio (music, sound effects, alarms) is NOT transcribed.",
    "confidence": 0.954,
    "uncertainty": "Whisper confidence is a no_speech_prob proxy; transcript content and alignment remain model-dependent."
  }
}

Question and answer

Question
What exactly is said starting about 4 seconds into this clip?
Answer
The transcript has the speaker saying "What you see here are the previous version and the new modified version side by side." from 3.8 seconds.
Check passedcontains all 3 phrases

How it was checked

Truth from
captions
Fixed at

Taken from the human captions published with the source: the caption line at 86.77 to 91.85 s in the source, which is 4.00 to 9.08 s in the clip. The clip is the source from 82.77 to 112.77 s, cut with no edit. The line was picked by a seeded rule (seed from the card id) from 21 eligible caption lines of 8 to 25 words, 2 to 10 s long, starting at least 4 s in, ending 0.5 s before the window does, with at least two words of 5 or more letters, and with no sound tag, music mark, dialogue dash or speaker label; the required phrases are the first 3 of those words in the line. No picture or sound was changed, and no tool was asked for the answer.

The check is the published caption file, and the 30-second clip was cut from the source with no edit, so a word that transcribe hears differently counts as a miss.

Try it yourself

Use agentic-video to transcribe my video, then quote the exact words said starting about 4 seconds in. Use my own file.
  1. Get the .mcpb file and open it in Claude Desktop.
  2. Or: pip install agentic-video

Clip and licence

Licence
CC-BY-3.0
Author
Living Colour Filmproduktion, Hamburg, for Wikimedia Deutschland
Source
commons.wikimedia.org
This file is licensed under the Creative Commons Attribution 3.0 Unported license.