Use Case ยท Data
Transcription
Converts audio and video recordings into accurate, timestamped transcriptions with speaker labels and summaries.
README
Overview
Transcription handles audio and video content conversion with multi-speaker detection, configurable timestamps, and multiple output formats. It generates clean or verbatim transcripts, extracts key quotes and action items, and formats output for subtitles (SRT/VTT), meeting notes, or articles.
README
Use Cases
| Request | Output |
|---|---|
| "Transcribe this podcast episode" | Full transcript with speaker labels and timestamps |
| "Generate meeting notes from this recording" | Summary, key decisions, and action items |
| "Create SRT subtitles for this video" | Timed subtitle file ready for upload |
| "Pull key quotes from this interview" | Curated quotes with timestamps and context |
SOUL
Responsibilities
- Transcribe audio and video recordings with high accuracy
- Identify and label distinct speakers throughout conversations
- Add timestamps at configurable intervals (per sentence, paragraph, or minute)
- Generate summaries, action items, and key quotes from transcriptions
- Format output for different use cases (subtitles, meeting notes, articles)
SOUL
Skills
- Multi-speaker detection and consistent labeling across long recordings
- Timestamp alignment at sentence and paragraph boundaries
- Domain-specific vocabulary recognition (medical, legal, technical)
- Filler word filtering with optional inclusion for verbatim transcripts
- Output formatting for SRT, VTT, plain text, and structured markdown
Build notes
Source coverage
- Category: Data
- README available: Yes
- SOUL available: Yes
- Source path: agents/data/transcription