| name | watch |
| description | Watch and understand footage with Diffusion Studio via the `dapi` CLI: answer questions about a video or audio file, summarize it, find scenes and moments, pull quotes, and describe what happens and when. Use whenever the user asks what's in a piece of footage, wants a summary or recap, wants to locate a moment ("where does X happen", "find the scene where..."), or needs a claim about a video or audio file checked. |
The CLI is self-describing and ships its own API reference. Use dapi --help, dapi media --help, and dapi media <command> --help to enumerate every command, argument, and option, and treat live help as authoritative rather than working from memory. If dapi is unavailable, read installation.md.
Watching footage
How to understand source material. Inspect only the modalities the question turns on — speech, action, music, graphics, or atmosphere may lead, so there is no fixed priority. Sample the picture against what the audio tells you.
- Always probe first.
dapi media probe <id|path> reports the container and its tracks, telling you up front whether the file has a video track, an audio track, or both. Everything after branches on that.
- Get the lay of the land. Render a
dapi media waveform (audio) and a dapi media filmstrip (video) for a fast, cheap overview of where the loud and quiet stretches fall, and where the visual scene changes are. A filmstrip shows coarse structure and scene state, not crop, framing, readability, or an exact cut frame.
- Listen to the audio. Run
dapi media listen with a prompt tailored to the context (what you actually need to know), and explicitly ask the model to include timestamps in its answer. See media-listen.md for prompt patterns.
- Transcribe speech. For speech,
dapi media transcribe prints the full transcript with word-level start/end times directly — read any segment straight from it.
- Sample the video against the audio. Use
dapi media grab to pull frames. When the audio has already pointed you at specific moments, feed those timestamps straight in from the transcript or listen output, e.g. -t '00:32' '00:45' .... When you need a visual pass without such cues, reach for --auto: it scans the footage and keeps only the frames where the picture settles into a new visual state, dropping near-duplicates.
Matching depth to the question
Read only as much of the footage as the answer requires — each pass costs time, and listen costs credits.
- A duration or format question ends at
probe.
- "Where is the quiet part" or "how is it paced" is usually answered by the waveform and filmstrip alone.
- A question about what was said resolves fastest through
transcribe; quote the transcript and its times directly.
- Questions about non-speech audio (music, tone, sound events, speaker identity) are what
listen is for.
- Only questions about what is seen need frames — and the audio pass usually tells you which moments to grab, so grab those instead of scanning blind.
- For open-ended questions ("summarize this", "what happens here"), combine passes: structure from waveform + filmstrip, content from transcript or listen, then frames at the salient moments to confirm what the picture shows.
Answering
- Ground every claim in something you actually saw or heard — a transcript line, a listen answer, a grabbed frame. If the evidence is ambiguous, say so rather than smoothing over it.
- Anchor answers to the timeline. Give timestamps as
MM:SS (or HH:MM:SS for long footage) so the user can jump straight to the moment; for a scene or segment, give its start and end.
- When asked to find a scene or moment, return the timestamp range plus a one-line description of what identifies it, so the user can confirm it is the right one.
- Summaries follow the footage's own structure: what happens, in order, with the timestamps where each part begins. Length matches what the user asked for, not what the footage contains.
- This skill only reads footage. When the user wants the footage changed — cut, composed, captioned, exported — that is the
editor skill's job.
