FAQ
What is loci?
An intelligence substrate for working with AI: plain-text templates and processes that decide how memory, context, and trust work, whichever AI runs them. Your setup is called a palace: a folder of markdown files you own. No cloud, no accounts, no lock-in.
Does loci send data anywhere?
No. A palace is plain text on your machine: nothing syncs, nothing phones home, and the CLI makes no network calls.
If you point a cloud-based AI (Claude, ChatGPT) at your palace, that provider processes what it reads, as with anything you give it. loci itself transmits nothing.
How do I run it?
Two ways, same substrate underneath:
- Plain markdown (the beginner path). Clone the repo, copy the
templates/kit into a folder of your own, openFIRST-SESSION.md, and point any file-aware AI at your palaceCLAUDE.md. No build. - CLI (
loci-cli, Rust, v0.6.0-beta). Five read-only commands, no network, no inference. See the CLI guide.
You do not need the CLI to start. The markdown kit is the whole substrate.
What AI tools are supported?
Any AI that can read files: Claude, Claude Code, Cursor, Codex, or a local model.
AGENT-SETUP.md in the repo has wiring notes for Claude Code, Cowork, and the Claude web interface. The palace moves between them unchanged.
Does it work with local models?
Yes. Any LLM with file access can run the substrate; nothing in loci assumes a provider. The CLI does no inference at all.
Is there an MCP server?
No. The archived desktop app carried one, and it went with the app. The CLI has no equivalent (loci serve is not shipped). You don't need one: a palace is plain text, so file-aware agents (Claude Code, Cursor) read it directly.
What is a crystal?
A unit of persistent memory with a tier and a lifecycle. Three tiers: ◇ exploratory, ◈ contextual, ◆ confirmed. You can promote, expire, pin, and compost them; time-sensitive crystals carry a valid_until date.
Keep what surprised you, and let the rest decay on its tier's own clock. More: Crystals, and templates/memory-lifecycle.md in the repo.
What is the difference between a room and a project?
A room is a persistent context workspace; a project is what you are building. One room can hold several projects, and one project can span several rooms. Rooms map to how you think: if you switch between "coding mode" and "writing mode", those are different rooms, even for the same project.
Is loci the same as a RAG system?
No. RAG retrieves text chunks and injects them into prompts. loci is the layer underneath: it decides what persists, how it is structured, and what loads when. Because a palace is plain text, you can still point any retrieval tooling at it.
Is there a web interface?
No. The surfaces are the plain files and the CLI. Nothing requires an account or a server.
Can teams share a palace?
Yes. The Team is one of the six use-case shapes in the repo's feature map. A palace is a folder of text files, so share it the way you share any text: a git repository or a shared folder. The Trust & Governance feature set exists to keep a shared palace honest.
How do I back up my data?
Back up the palace folder like any folder: rsync, Time Machine, git, whatever you use. If you ran loci init, the CLI also keeps a config at ~/.config/loci/config.toml.
How do I keep my palace up to date?
Ask your AI ("Update my palace") and point it at the repo's PALACE-UPDATE.md. It compares your palace with the published methodology version and offers the new pieces one at a time. Nothing is auto-applied.
What happened to the browser extension, the tiers, and the desktop app?
Earlier versions of these docs described a browser extension, a conversation index, and three product tiers. None of that is the shipped product (the correction). loci.garden keeps two looks, wizard and llmage; they style the site, not the product.
The current release is v0.6.0-beta (2026-07-02). A desktop app shipped in that beta and has since been archived: loci runs as the markdown kit plus the optional CLI. Where older pages disagree, trust this one and the repo's CHANGELOG.md.