Grimoire FAQ

Can I sync my bookmarks across devices?

Not yet. Grimoire is local-first and single-user for 1.0.1. Backups are snapshot-based, so you can move or restore a saved backup, but there is no continuous multi-device sync service.

Is my data private?

Yes, by default. Bookmarks, extracted content, notes, tags, categories, embeddings, settings, backups, and logs are stored locally. External AI providers are optional; if you configure one, the content sent to that provider is governed by that provider's terms. You can use Grimoire without AI providers.

What happens if the daemon crashes?

Bookmarks and jobs are stored in SQLite with WAL mode. Restart the daemon with your platform service manager or run npm run daemon:dev again during development. The durable job queue resumes pending ingestion work after the daemon starts.

Can I use it without AI providers?

Yes. Saving, importing, keyword search, tags, categories, archive/trash, backup/restore, diagnostics, and local integrations work without AI. The UI shows degraded-mode guidance when summaries, generated tags, semantic search, or embeddings need a configured provider.

How do I upgrade?

For source or unpacked release installs, run daemon/install.sh --upgrade. Packaged installs can use littleimp update install once a reachable release source is configured. Settings can also check for update availability. See update-system.md.

Does it work on Windows?

There is no native Windows installer in 1.0.1. Docker on Windows with WSL2 is the supported path. Keep Docker port publishing bound to loopback, for example 127.0.0.1:3210:3210.

Can I run it on a server?

Yes through Docker, but Grimoire itself is not a public-server product in 1.0.1. Keep the daemon loopback-bound or put an authenticated tunnel, VPN, or reverse proxy in front of it before traffic reaches Grimoire. See security-boundaries.md and docker-deployment.md.