Linggen Local RAG
- Updated July 27, 2026
- linggen/linggen-releases
Linggen is an MCP server that connects coding agents to a local, privacy-focused RAG service for searching your codebase.
About
Linggen is a Model Context Protocol server that fronts a local retrieval-augmented generation service tuned for searching code. Developers who ship with Claude Code, Cursor, or similar agents install it when they want semantic or hybrid search across repositories without routing source through a hosted code-intelligence vendor. The catalog positions it in the Build phase under integrations because the payoff appears while you are extending an agent’s toolchain and wiring context providers, not when you are validating a landing page or tuning SEO. Compared to cloud-only codebase assistants, Linggen emphasizes on-device indexing and query paths so sensitive forks, client work, or air-gapped environments remain viable. You register the server in your MCP config, point it at your local Linggen service, and the agent can pull relevant snippets during implementation, refactors, and incident triage. It is an integration layer, not a replacement for git grep or IDE navigation; expect to run and maintain the Linggen runtime alongside the MCP shim described in the official release metadata on GitHub.
- Runs locally for privacy-first RAG over code without mandatory cloud ingestion
- Exposes code search to MCP clients (Claude Desktop, Claude Code, Cursor)
- Targets builder workflows documented at linggen.dev with server schema version 0.1.2
- stdio MCP transport pattern for agent-side registration
- Pairs with agent builds where context must stay on-machine
Linggen Local RAG by the numbers
- Data as of Jul 28, 2026 (Skillselion catalog sync)
claude mcp add Linggen-LocalRAG -- npx -y linggen/linggen-releasesAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| Last updated | July 27, 2026 |
|---|---|
| Repository | linggen/linggen-releases ↗ |
What it does
Give Claude Code or Cursor a local, private retrieval layer so the agent can search your repos without sending code to a cloud indexer.
Who is it for?
Best when you're on Claude Code or Cursor and need agent-grounded code retrieval in private repos, client code, or offline-friendly setups.
Skip if: Skip if you want fully managed cloud semantic search with zero local ops, or developers and only need simple ripgrep without embeddings.
What you get
After you register Linggen in MCP, your agent can query local code context through RAG-style search while keeping indexing under your control.
- MCP-registered bridge from your agent to local code RAG queries
- Agent-accessible code search without mandatory cloud ingestion
By the numbers
- Published MCP server schema version 0.1.2 in server metadata
- Repository source: github.com/linggen/linggen-releases
README.md
Linggen Releases
Public download host for branded Linggen apps on macOS. All published builds are codesigned and notarized.
Linggen is a local AI app engine — the core runtime (engine source) hosts skills as full-featured native apps. Each branded app bundles the engine + a primary skill into one
.app. Apps share a local daemon (http://127.0.0.1:9898) and the~/.linggen/data directory.
Apps
Sys Doctor — coming soon
On-device macOS health analyst. Inspects disk, memory, processes, network, security, and battery; explains findings in plain English; suggests cleanup with copy-paste commands.
- Tag scheme (when released):
sys-doctor-v<version> - Source: linggen/linggen-app (private)
More branded apps planned (House Keeper, Majiang, Suanming, …) — see the roadmap.
Install (once a DMG is published)
- Download the
.dmgfor the app you want from Releases. - Open it; drag the app into
/Applications. - First launch: the app spawns the bundled
lingengine on127.0.0.1:9898(or reuses one already running). The daemon self-terminates after 5 min idle.
All Linggen apps share one local daemon and one ~/.linggen/ directory (sessions, memory, skills, credentials).
Tag schemes
| Tag pattern | Where | What |
|---|---|---|
<app>-v<version> |
this repo | branded native app .dmg (e.g. sys-doctor-v0.1.0) |
v<version> |
linggen/linggen |
the ling CLI / engine binary on its own |
The engine binary is not published here. Branded apps bundle their own copy of ling; this repo only hosts the user-facing app DMGs.
Recommended MCP Servers
0Latency MemoryPersistent memory layer for AI agents.
0nMCP — Universal AI API OrchestratorUniversal AI API Orchestrator — 1,554 tools, 96 services.
0xHumans Protocol MCPMCP for AI agents: financing, skills, lending on Base1k Patient Mcp1k patient MCP server
1trippulse1trip PULSE: 21-tool AI travel planner.
3D AI Agent Avatar3D AI Agent Avatar — render any GLB, give it a Solana wallet, a voice, and pump.fun powers.89How it compares
MCP integration to a local RAG runtime, not a hosted IDE plugin or a single prompt skill.
FAQ
Who is Linggen for?
Linggen is for developers and small teams using MCP-enabled agents who want codebase search backed by local RAG instead of sending source to external indexers.
When should I use Linggen?
Use Linggen during active development when you are integrating agent tooling and need repeated, semantic code lookup across large or multi-repo projects.
How do I add Linggen to my agent?
Install and run the Linggen local service per linggen.dev, then add the io.github.linggen/linggen MCP server entry to your client’s MCP configuration so tools proxy search to that service.