How is this different from Cursor’s own memory features?

Cursor’s context lives inside your editor and repo. LUCI’s memory covers everything outside it — browser, terminal, other apps, meetings — and it isn’t vendor-scoped: the same official skills work in Claude Code and Codex too.

Project-level or global install?

From LUCI, the skills land in ~/.cursor/skills/ and apply everywhere. From the repo, `npx skills add Memories-ai-labs/Luci-skills` can be per-project, or add `-g` for every project. Optional MCP config works the same way: global ~/.cursor/mcp.json or a project .cursor/mcp.json.

Do I need to keep LUCI running?

For live history, yes. If LUCI is quit, the luci skill reads your Life folder instead. The memory only grows while LUCI is capturing.

Coding agent

Give Cursor Memory Beyond the Codebase

Cursor knows your repo. It doesn’t know the error you saw in the browser, the doc you read at lunch, or what your team decided on the call. Install the official LUCI skills and Cursor’s agent can search all of it — from a memory stored entirely on your machine.

See all integrations

Cursor’s agent is deeply aware of your codebase — and completely unaware of everything that happens around it. Debugging lives in browser tabs, dashboards, terminal windows, and meetings; when you ask Cursor for help, you become the courier who ferries that context in by hand.

LUCI turns that surrounding context into something Cursor can query. Install the official skills from the app or from Memories-ai-labs/Luci-skills. MCP is available if you want the tools without the skill files.

What Cursor can do with your memory

Debug with what you actually saw

The exception in the browser console, the failing request in the network tab, the log line from a dashboard — Cursor can search the text of all of it instead of asking you to paste it.

Bring meetings into the editor

Locally transcribed mic and system audio means Cursor can pull up what was decided in the planning call while you implement it.

Trace your own steps

Filter captures by app and time, or aggregate a whole range — "what was I doing before this broke?" becomes a query, not an archaeology session.

Keep it all on-device

Capture, understanding, and storage happen on your machine, encrypted at rest. Agents ask for what they need, and only that much enters the conversation.

Try asking

Questions that work on day one

Search my screen memory for the TypeError I saw in the browser console
What was on my screen when the deploy failed?
Find where we discussed the schema change in today’s call
List what I had open in Chrome this morning

Set up Cursor

01
Turn on Screen Memory in LUCI
Install LUCI on your Mac or Windows machine and enable Screen Memory. Leave the app running.
02
Install the official skills
In LUCI, open Settings → Agent access, tick Cursor, and click Install. Or install from the official repo: https://github.com/Memories-ai-labs/Luci-skills
```
npx skills add Memories-ai-labs/Luci-skills
```
03
Optional: connect over MCP
If you want Cursor to call LUCI as a tool without the skills, add the ready-made snippet to ~/.cursor/mcp.json (or a project-level .cursor/mcp.json).
```
{
  "mcpServers": {
    "luci": {
      "url": "<url>",
      "headers": { "Authorization": "Bearer <token>" }
    }
  }
}
```

Where Cursor looks for it

The two skill folders are ordinary files you own. Drop them where this tool reads and it finds them on the next session.

Skills folder: ~/.cursor/skills/, or .cursor/skills/ per project
Written by the LUCI installer: Yes, Settings > Agent access
Optional MCP file: ~/.cursor/mcp.json, or .cursor/mcp.json per project

Frequently asked questions

The memory itself stays on your machine. What an agent actually pulls becomes part of that conversation. If the agent runs in someone else’s cloud, that conversation goes there. Connect only the agents you trust, and ask them for what you are willing to put in that thread.