Research Assistant

Process Bear notes tagged 「待整理」: extract topics, find matching GIFs, insert them, then remove the tag.

Workflow

1. Fetch pending notes — run scripts/process_notes.sh list to get all notes with the 待整理 tag.
2. For each note, run scripts/enrich_note.sh which:
• Reads the note content via grizzly open-note
• Extracts key topic phrases (first H1, bold terms, or first sentence)
• Searches for a relevant GIF via Tenor API (falls back to Giphy)
• Appends the GIF as a markdown image at the end of the note
• Removes the 待整理 tag and adds 已整理 tag
3. Report — summarize how many notes were enriched and any failures.

Requirements

• Bear app running on macOS
• grizzly CLI installed (go install github.com/tylerwince/grizzly/cmd/grizzly@latest)
• Bear token at ~/.config/grizzly/token
• TENOR_API_KEY or GIPHY_API_KEY env var (free tier works)

Scripts

• scripts/process_notes.sh list — list note IDs with 待整理 tag
• scripts/enrich_note.sh — enrich a single note with a GIF