Skip to main content
Glama
jamalhansen

local-first-mcp

by jamalhansen

local-first-mcp

MCP server that exposes local-first AI tools to Claude.

This server allows Claude to use your locally-built tools (like resource-summarizer, promo-generator, etc.) directly via the Model Context Protocol.

Tools Exposed

  • summarize_article: Fetch and summarize a URL (via resource-summarizer)

  • generate_promo: Create promotional copy for a blog post (via promo-generator)

  • semantic_search: Search your Obsidian vault by meaning (via vault-semantic-search)

Related MCP server: Obsidian MCP Server

Installation

cd local-first-mcp
uv sync

Configuration for Claude Desktop

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "local-first": {
      "command": "uv",
      "args": [
        "--directory",
        "/Users/jamalhansen/projects/local-first/local-first-mcp",
        "run",
        "local-first-mcp"
      ]
    }
  }
}

Project Structure

local-first-mcp/
├── src/
│   └── local_first_mcp/
│       ├── server.py    # FastMCP server definition
│       └── tools/       # Tool-specific wrappers
│           ├── promo.py
│           ├── search.py
│           └── summarize.py
├── pyproject.toml       # Managed by uv
└── tests/               # Server integration tests

Related MCP Connectors

Related MCP Servers

  • F
    license
    Not graded
    quality
    D
    maintenance
    Provides semantic search capability over Obsidian vaults and exposes recent notes as resources to Claude through the MCP protocol.
    9
    -
  • A
    license
    Not graded
    quality
    B
    maintenance
    Provides Claude with tools to read/write an Obsidian vault, query a local SQLite database, and manage files, accessible from any device via OAuth over Tailscale.
    MIT
  • A
    license
    B
    quality
    D
    maintenance
    Enables Claude to interact with an Obsidian vault via the Local REST API plugin, supporting file operations, search, periodic notes, commands, and semantic search.
    22
    2,509 npm
    MIT