mnueron — VS Code extension

Save, search, and inject AI memories from inside VS Code. Works alongside GitHub Copilot, Cursor, Cline, or any other AI assistant — mnueron is the memory, not the AI.

Already using Cursor or Cline? Those AIs read mnueron natively via MCP — this extension is complementary, giving you a sidebar to browse memories and shortcuts to save selections. Useful even when your AI already has memory access.

Using GitHub Copilot? Copilot doesn't speak MCP. This extension is how you give Copilot users mnueron — they can hit Cmd+Shift+R to recall context, paste it into the Copilot chat, and have it use the recalled memory.

What you get

Sidebar memory list (activity bar → mnueron). Shows your 25 most recent memories in the active namespace. Click any to open in a peek view.
Save selection — Cmd+Shift+M (mac) / Ctrl+Shift+M (win). Saves the current editor selection as a memory, tagged with the file language. With no selection, prompts you for content via input box.
Recall — Cmd+Shift+R / Ctrl+Shift+R. QuickPick search; picking a result inserts it at the cursor (or copies to clipboard with no editor).
Namespace switcher in the status bar.
Ambient context (optional): when you open a file, mnueron quietly searches for related memories and shows a count in the status bar.

Setup

Local mode (free, no account)

Install the mnueron CLI:
```
npm install -g mnueron
```
Start the local dashboard server in a terminal (or via the CLI's setup wizard):
```
mnueron dashboard
```
In VS Code Settings → mnueron, leave mnueron.mode as local.

The extension talks to 127.0.0.1:3122 (no auth needed) — all your memories stay on your machine at ~/.mnueron/memories.db.

Hosted mode (sync across machines)

Sign up at https://www.mnueron.com.
Issue a token at /account-settings/tokens (copy the mnu_... string).
In VS Code Settings → mnueron:
- Set mnueron.mode to hosted.
- Paste the token into mnueron.apiToken.

Configuration

Setting	Default	Description
`mnueron.mode`	`local`	`local` (CLI dashboard) or `hosted` (mnueron.com)
`mnueron.hostedUrl`	`https://www.mnueron.com`	Override for self-hosting
`mnueron.localUrl`	`http://127.0.0.1:3122`	Local CLI dashboard URL
`mnueron.apiToken`	`""`	Bearer token (hosted mode only)
`mnueron.activeNamespace`	`vscode`	Namespace for saves from this editor
`mnueron.ambientContext`	`false`	Search for related memories when opening files

Workflow examples

Capture a design decision while you're editing:

Select the relevant code lines.
Hit Cmd+Shift+M (or right-click → "mnueron: Save selection as memory").
Memory is saved with file path + language metadata; you can recall it later by file name or by content.

Recall context before asking Copilot a question:

Hit Cmd+Shift+R.
Type a few words ("auth approach", "rate limit").
Pick a result; it's inserted at the cursor (paste into Copilot's chat panel as context).

Pull yesterday's notes into a fresh planning doc:

Open Cmd+Shift+P → "mnueron: Open dashboard in browser".
From the dashboard, use the Context Builder to multi-select and copy-as-prompt.

Building from source

git clone https://github.com/randi2160/mnueron.git
cd mnueron/sdks/vscode
npm install
npm run build

Then either:

Run a dev instance: Open the folder in VS Code, press F5 to launch an Extension Development Host.
Build a VSIX: npx vsce package → install the resulting .vsix via "Install from VSIX…" in VS Code's command palette.

What's NOT here yet

No automatic git-commit capture (planned for v0.2)
No CodeLens "you wrote N memories about this file" (planned for v0.2)
No marketplace publication yet — install via VSIX for now
No remote dev container support (works locally only)

PRs welcome.

License

MIT.

mnueron — AI memory

mnueron