Linggen

Linggen is a local-first memory layer for AI coding assistants that runs entirely on your machine. It provides a shared memory space across projects, enabling fast semantic search and AI-enabled chat over your indexed code, docs, and notes, while exposing an MCP server for integration with your IDEs and agents. All embeddings, indexes, and search stay local by default for privacy and performance.

How to use

You connect your MCP client to Linggen’s server at http://localhost:8787/mcp/sse to enable cross-IDE memory access and collaboration. Use Linggen to index your folders, codebases, documentation, and notes, then perform semantic search or chat over everything you’ve indexed. Point your MCP client to the Linggen server so you can navigate large codebases and share memory across tools and teams.

How to install

Prerequisites: ensure you have a supported operating environment for Linggen. The current build is distributed as a macOS application package.

1. Download the latest macOS build from the official releases page.

2. Open the downloaded .dmg file and drag Linggen into the Applications folder.

3. Launch Linggen. On first run, it will download an embedding model and start a local backend at http://localhost:8787.

Note: Windows and Linux builds are planned for future releases.

Additional sections

MCP server configuration: you can connect clients like Cursor to Linggen’s MCP endpoint. The server URL to use is http://localhost:8787/mcp/sse. You will typically add this URL to your MCP client’s configuration under the servers list.

{
  "mcpServers": {
    "linggen": {
      "url": "http://localhost:8787/mcp/sse"
    }
  }
}

Available tools

semanticSearch

Performs semantic search and AI-assisted querying over all indexed data, enabling fast, relevant results across code, docs, and notes.

mcpExposure

Exposes an MCP server endpoint so compatible clients can connect and share memory across tools and IDEs.

worldIndexing

Index local folders, codebases, docs, and notes to build a searchable memory of your entire workspace.