Installable skillApache-2.0

case.dev Transcription

Transcribes audio and video files through case.dev with speaker diarization. Supports MP3, WAV, M4A, FLAC, OGG, WEBM, MP4 up to 5GB. Use when the user mentions "transcribe", "transcription", "deposition recording", "hearing audio", "speaker labels", or needs to convert audio/video to text.

BY CaseMarkUPDATED Oct 28, 2026
INTENDED FOR✳Claude Code>_Codex◈OpenClaw↗CursorInstall in a listed agent.

instructions.md

---
name: transcription-bed905528e25
language: en
description: Transcribes audio and video files through case.dev with speaker diarization. Supports MP3, WAV, M4A, FLAC, OGG, WEBM, MP4 up to 5GB. Use when the user mentions "transcribe", "transcription", "deposition recording", "hearing audio", "speaker labels", or needs to convert audio/video to text.
---

# case.dev Transcription

Audio and video transcription with speaker diarization, ideal for depositions, hearings, and recorded proceedings. Supports MP3, WAV, M4A, FLAC, OGG, WEBM, and MP4 up to 5GB.

Requires the `casedev` CLI. See `setup` skill for installation and auth.

## Start a Transcription

The source file must be in a vault:

```bash
casedev transcribe run --vault VAULT_ID --object OBJECT_ID --json
```

Flags: `--speaker-labels` (diarization), `--speakers-expected N`, `--language`, `--format`, `--auto-highlights`.

## Check Status

```bash
casedev transcribe status TRANSCRIPTION_ID --json
```

Statuses: `queued` -> `processing` -> `completed` or `failed`.

## Get Result

```bash
casedev transcribe result TRANSCRIPTION_ID --json
```

Fetches transcript text. Fails if the job is not yet completed.

## Watch Until Complete

```bash
casedev transcribe watch TRANSCRIPTION_ID --json
```

Flags: `--interval` (default: 3s), `--timeout` (default: 900s).

## Common Workflow

```bash
# 1. Upload to vault
casedev vault object upload ./deposition-recording.mp3 --vault VAULT_ID --json

# 2. Start transcription with speaker labels
casedev transcribe run --vault VAULT_ID --object OBJECT_ID --speaker-labels --speakers-expected 4 --json

# 3. Watch until complete
casedev transcribe watch TRANSCRIPTION_ID --json
```

## Troubleshooting

**"Invalid file type for transcribe"**: Only audio/video MIME types accepted. Check with `casedev vault object list`.

**Transcription failed**: Check `casedev transcribe status TRANSCRIPTION_ID --json` for error details. Common causes: corrupted file, unsupported codec, file too large.

**Long audio (2+ hours)**: Increase timeout with `--timeout 3600`.
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.