Linggen — Portable AI Memory for Your Editor
Bring ling-mem — portable, user-owned AI memory — into VS Code, Cursor,
Windsurf, Codex, and Copilot. Your preferences, decisions, and gotchas follow
you across every AI tool you use.
This extension is a thin installer and bridge. It installs the ling-mem
binary, runs its local dashboard, and wires it into the AI agents you already
chat with. There is no new chat to learn — the agents in your editor become
memory-aware on their own.
ling-mem currently ships macOS and Linux binaries only. Windows support
is not available yet. The extension will not function on Windows until
ling-mem adds a Windows build.
What it does
On first activation:
- Downloads the
ling-mem release pinned by install-ling-mem.sh (macOS / Linux)
- Installs the ling-mem skill bundle for Claude Code (and your project's
CLAUDE.md, Cursor rules, AGENTS.md, etc. — whichever agents you opt into)
- Asks which AI agents to bridge (Claude Code, Cursor, Codex, Copilot,
Windsurf, Zed)
- Injects ling-mem instructions into the chosen agent rule files, inside a
<!-- ling-mem:start -->...<!-- ling-mem:end --> block
- Starts the ling-mem local daemon
On every activation:
- Re-runs
install-ling-mem.sh if the binary is missing
- Ensures the daemon is running (uses an already-running daemon's port if one
exists)
- Reconciles the delimited bridge sections in current-workspace rule files
to match your consent (writes for opted-in targets, strips from opted-out
targets)
At runtime:
- Status bar shows
Linggen ✓ / … / ✗
- Command
Linggen: Open Dashboard opens ling-mem's dashboard in the browser
- Command
Linggen: Bridge Settings re-runs the consent picker
Want automated memory maintenance? Try Linggen
ling-mem ships a dream mission — a nightly job that consolidates and
deduplicates your memory, merges near-duplicates, and prunes stale rows. The
mission file is installed at ~/.linggen/missions/dream/mission.md and is
scheduled for 23:00 local time. It runs entirely on your machine.
The mission file is defined by ling-mem, but the scheduler that actually
fires it lives in Linggen — the local AI app engine
that hosts ling-mem and other skills. Without Linggen, the dream mission sits
on disk but never runs on its own.
If you want it to run automatically — and you're open to running a local AI
engine that can also manage other agents, missions, and skills — install
Linggen:
curl -fsSL https://linggen.dev/install.sh | bash
ling-mem itself works fine without Linggen for everyday capture and recall;
Linggen is the upgrade path when you want memory maintenance running on its
own schedule.
Configuration
| Setting |
Default |
Description |
linggen.dashboard.host |
127.0.0.1 |
Host for the ling-mem dashboard |
linggen.dashboard.port |
9888 |
Port for the ling-mem dashboard |
Telemetry
Anonymous install + command events are sent to linggen.dev to measure usage.
Opt-out by setting LING_MEM_NO_TELEMETRY=1 or creating
~/.linggen/no-telemetry. The ling-mem skill bundle never phones home — only
the extension and the ling-mem binary do.
Links