Grov

Grov captures the reasoning and decisions your team's AI agents discover and syncs them to a shared team dashboard. When one developer's Claude figures something out—like how your authentication system works—every other developer's Claude automatically knows it without re-exploring. The tool injects relevant team memories into new sessions, eliminating redundant investigation and token spend. It's designed for engineering teams using Claude Code, Cursor, or other Claude-powered tools who want their AI agents to learn collectively rather than in isolated silos.

1. **Start a Private Session:** Open your AI tool (e.g., Claude, ChatGPT) and begin a private session focused on a specific task or problem. Include detailed context like code snippets, error logs, or project goals. 2. **Trigger Grov Capture:** After resolving a subtask or reaching a decision, use the prompt template above to generate a Grov-compatible summary. Replace [SESSION_TRANSCRIPT] with your session notes or copy the entire transcript. 3. **Sync to Team Memory:** Paste the formatted output into your team's Grov channel (e.g., Slack, Discord, or Grov's web interface). Ensure you include relevant tags (#backend, #priority-high) to categorize the entry for easy retrieval. 4. **Auto-Injection in Future Sessions:** In subsequent private sessions, Grov will automatically inject relevant memories (e.g., previous decisions, unresolved questions) to provide context and save tokens. For example, if you’re debugging again, Grov might surface the `timeoutMiddleware` fix from the shared memory. 5. **Refine and Iterate:** After syncing, review the shared memory entry for accuracy. If Grov misses critical context, manually edit the entry or provide additional details in your next session to improve future auto-injections.