AI Editor Integrations (MCP)

This page documents practical MCP integration patterns for the most common AI coding environments.

Recommended Ecommaps Transport

• Primary: GET/POST https://api.ecommaps.com/api/v1/mcp
• Auth: Authorization: Bearer <ECOMMAPS_API_KEY>

Prefer Streamable HTTP transport by default for easier debugging, gateway policy enforcement, and production observability.

Supported Editor Ecosystem

1. Visual Studio Code (Copilot + MCP support)
   Reference: VS Code MCP documentation (opens in a new tab)

2. Cursor
   Reference: Cursor MCP documentation (opens in a new tab)

3. Claude Code (Anthropic)
   Reference: Claude Code MCP documentation (opens in a new tab)

4. Windsurf
   Reference: Windsurf MCP documentation (opens in a new tab)

5. GitHub Copilot coding agent with MCP servers
   Reference: GitHub MCP integration docs (opens in a new tab)

Integration Checklist

1. Create a dedicated API key per tool/runtime from store settings.
2. Label keys by environment and owner (cursor-dev, vscode-team, claude-prod).
3. Configure MCP client endpoint to Ecommaps Streamable HTTP URL (/api/v1/mcp).
4. Validate tool discovery (tools/list) before enabling agent tasks.
5. Audit and rotate keys regularly.

Security Notes

• Never store API keys in editor settings that sync to public profiles.
• Prefer environment variable injection for local and CI usage.
• Revoke keys immediately if leaked in logs or prompts.