Connect to a Coding Agent

By the end of this guide, your code editor's AI assistant will be able to search your team's indexed knowledge — Jira tickets, Confluence pages, and local docs — right from your editor.

Prerequisites

• Indexed installed
• At least one collection created (see Quick Start or any connector guide)

How It Works

Indexed runs as an MCP server using stdio transport — your editor launches it as a subprocess and communicates via stdin/stdout. The config format is nearly identical across all tools.

Cursor

Create .cursor/mcp.json in your project root (project-scoped) or ~/.cursor/mcp.json (global):

{
  "mcpServers": {
    "indexed": {
      "command": "indexed",
      "args": ["mcp", "run"]
    }
  }
}

Cursor will pick up the config automatically. Open Agent mode and ask something that should pull from your docs.

Windsurf

Edit ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "indexed": {
      "command": "indexed",
      "args": ["mcp", "run"]
    }
  }
}

GitHub Copilot (VS Code)

Create .vscode/mcp.json in your project root (or use the command palette: MCP: Open User Configuration for a global config):

{
  "servers": {
    "indexed": {
      "type": "stdio",
      "command": "indexed",
      "args": ["mcp", "run"]
    }
  }
}

Note: GitHub Copilot uses servers (not mcpServers) and requires an explicit "type": "stdio".

Cline (VS Code)

Open the Cline extension, click the MCP Servers icon, then Configure. Add to cline_mcp_settings.json:

{
  "indexed": {
    "command": "indexed",
    "args": ["mcp", "run"],
    "disabled": false
  }
}

Zed

Edit ~/.config/zed/settings.json:

{
  "context_servers": {
    "indexed": {
      "source": "custom",
      "command": "indexed",
      "args": ["mcp", "run"]
    }
  }
}

Continue.dev

Create a YAML file at .continue/mcpServers/indexed.yaml:

name: Indexed
version: 0.0.1
schema: v1
mcpServers:
  - name: indexed
    command: indexed
    args:
      - mcp
      - run

Troubleshooting

Tools not appearing

• Ensure indexed is on your PATH — run which indexed in a terminal
• Restart your editor after editing the config file
• Verify you have at least one collection: indexed index inspect

Slow first response

The first query loads the embedding model into memory. Subsequent queries are fast.

What's Next