product-os: Command for Claude Code

Export Granola Transcripts

Export meeting transcripts from Granola's local cache to markdown files. Useful for backing up transcripts, bulk searching across meetings, or working with transcripts offline.

Instructions

1. Check the auto-memory (MEMORY.md) for the user's initials (e.g. CR). Set {initials} accordingly.

2. Run the export script:

   python3 scripts/export_granola_transcripts.py
   

3. The script will:

   • Read Granola's local cache at ~/Library/Application Support/Granola/cache-v3.json
   • Parse all meetings and their transcripts
   • Create markdown files with meeting title, date, participants, speakers, and full transcript
   • Skip files that already exist (safe to re-run)
   • Output to team/{initials}/granola-transcripts/ (gitignored) by default

4. The output directory can be customised via the GRANOLA_BACKUP_DIR environment variable:

   GRANOLA_BACKUP_DIR=/path/to/dir python3 scripts/export_granola_transcripts.py
   

5. After export, report to the user:

   • How many new transcripts were exported
   • How many were skipped (already existed)
   • Date range covered
   • Where the files are located

6. If the user wants to search for something across their transcripts, use Grep to search the output directory.

Notes

• Granola must be installed and have cached meetings locally for this to work
• Transcripts are gitignored — they contain potentially sensitive meeting content
• File naming format: YYYY-MM-DD_HHMM_{safe_title}.md