Connect your tool

The tool you already run reads and writes the project over MCP. Copy its snippet from Integrations → AI tools in the app: that copy carries a key made for the tool, which reaches this project and nothing else.

Claude Code

claude mcp add --transport http runlog-memory https://runlogai.com/runlogapi/mcp/ext

Cursor and VS Code

mcp.json (Cursor, Windsurf, VS Code)

{
  "mcpServers": {
    "runlog-memory": {
      "url": "https://runlogai.com/runlogapi/mcp/ext"
    }
  }
}

ChatGPT and claude.ai

Add the address as a connector.

The address

https://runlogai.com/runlogapi/mcp/ext

Memory for your agents

Ask What does this repository do? The tool calls memory_query and answers from what the project holds, with its sources. What you tell it to remember lands through memory_write, where Ro and every other connected tool find it.

SeeTo keep a repository’s code current in the project, link the folder. Building your own agent: the API.