Recipe
Clip the useful answer without moving the whole recording to a hosted editor.
Manual clip rendering is for the moment after transcription: choose the source range, create an edit decision, render the artifact, and keep the working material in the instance you operate.
- Built-in recipe
- manual_clip_render
- Input
- Recording id plus source range or cue ids
- Output
- Local render manifest and MP4 artifact
Shell
manual-clip-render.sh
curl -sS -X POST "$VOXNOTCH_URL/workflow-recipes/manual_clip_render/run" \
-H "X-API-Key: $VOXNOTCH_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"recording_id": "rec_example",
"clip": {
"source_ranges": [
{ "start_ms": 65000, "end_ms": 95000, "label": "answer-that-matters" }
]
},
"render": {
"render_config": { "format": "mp4" }
},
"execute": true,
"idempotency_key": "answer-that-matters-v1"
}'Use a stable idempotency key for repeatable evaluation runs. Change it only when the source range or render intent changes.
Evaluation evidence
The evidence should show the path, not private content.
Recording
Capture the recording id and duration, with private transcript text redacted.
Range
Capture the start and end milliseconds or selected cue ids.
Artifact
Capture the local render artifact path, size, and created timestamp.