-
Notifications
You must be signed in to change notification settings - Fork 1
FAQ
No. Memora is 100% local. All data stays on your machine at ~/.memora/. There is no cloud sync, telemetry, or analytics.
Only for the Blackboard feature (which calls an LLM API). All other features work completely offline.
Memora will recreate the directory on next launch and re-import available history from your AI tools. However, hook event logs and some conversation history may be lost if the tools have already cleaned up their own data.
Not yet. Support for additional tools (Windsurf, Aider, etc.) is planned.
Not planned for now. Memora is designed for developers using heavy-duty AI coding agents (Claude Code, Codex, Cursor, OpenCode) — tools that generate long conversations, complex decision chains, and rich context that's painful to lose.
Lightweight AI editors (code completion, quick Q&A) don't produce the kind of deep context that needs cross-session memory management. If these tools evolve into full agent experiences, we'll reconsider.
- Check that hooks are installed: look for the hook entries in your tool's config
- Try having a conversation with your AI tool, then wait 15 seconds for auto-sync
- Click the refresh button in Memora's title bar
Right-click the app → Open → Open. This only needs to be done once. We're working on Apple notarization to eliminate this warning.
Check ~/.memora/.git/ size. If it's unexpectedly large, run:
rm -rf ~/.memora/.gitMemora will recreate a clean git repository on next sync.
Go to Settings → LLM Providers → Add, and configure at least one API key.
Yes. Copy ~/.memora/ to your backup location. The raw/ directory contains the original conversation transcripts (most valuable). memora.db is a cache that can be rebuilt.
Not currently. The data directory is hardcoded to ~/.memora/.
Only when you explicitly use Continue V2 (creates a new directory) or History Branching (creates a new branch). Memora never modifies your existing project code.