MCP Integration

ContextWeaver can run as a Model Context Protocol server, allowing MCP-compatible AI clients to retrieve context from the current codebase.

Start the MCP server

contextweaver mcp

Recommended workflow

1. Run contextweaver init in the target codebase.
2. Configure ~/.contextweaver/.env.
3. Run contextweaver index to build the initial index.
4. Configure the ContextWeaver server in your MCP client.
5. Use semantic retrieval, file listing, symbol definition, and reference tools to fetch code context.

Configure in your client

ContextWeaver runs over stdio with the command contextweaver mcp. Here is how to connect two common clients.

Codex

Codex stores MCP configuration in config.toml (default ~/.codex/config.toml, or .codex/config.toml in a trusted project).

Add it via the CLI:

codex mcp add contextweaver -- contextweaver mcp

Or edit config.toml directly:

[mcp_servers.contextweaver]
command = "contextweaver"
args = ["mcp"]

Run /mcp in the Codex TUI to see connected servers.

Claude Code

Claude Code adds stdio servers with claude mcp add. All options must come before the server name, and the command to launch follows --:

claude mcp add contextweaver -- contextweaver mcp

Use --scope to choose where the config lives: local (default, private to you in the current project), project (written to .mcp.json at the project root and shared via version control), or user (available across all your projects). Project-scoped example:

claude mcp add --scope project contextweaver -- contextweaver mcp

A project-scoped server is written to .mcp.json:

{
  "mcpServers": {
    "contextweaver": {
      "command": "contextweaver",
      "args": ["mcp"]
    }
  }
}

Run /mcp inside Claude Code to check the connection status.

If contextweaver is not on your PATH, replace command with the full path to the executable (find it with which contextweaver).

MCP tool capabilities

• Semantic code retrieval
• Indexed file listing
• Symbol definition lookup
• Symbol reference lookup
• Index and health statistics

On the first query, ContextWeaver can automatically trigger indexing. Later queries reuse local SQLite and LanceDB indexes.