Model Context Protocol (MCP)

Give your AI assistants, IDEs, and agents a single, secure, and persistent memory server using the open standard protocol.

Model Context Protocol

Model Context Protocol (MCP) is an open-source standard designed by Anthropic that allows local or remote servers to safely expose data tools and contexts to Large Language Models (LLMs) like Claude, GPT, and code assistants.

Rulip integrates natively as an MCP server. This means you do not need to install complex local databases or manually copy context files. Rulip spins up a secure SSE (Server-Sent Events) container and serves three simple, powerful tools to your AI:

  1. save_memory(content: str): Automatically saves context, links, summaries, and structural facts to your private memory cloud.
  2. search_memory(query: str): Instantly queries your past logs across chats and IDE workspaces for context injection.
  3. list_memories(limit: int): Retrieves the most recent logs to construct a rolling short-term window of previous work.

Select your preferred platform in the sidebar to get the configuration values and connect Rulip to your workflow in under 2 minutes.

Claude Claude Desktop

To connect Rulip memory to your Claude Desktop application, edit your local configuration file and add the Rulip MCP Server details.

  1. Open your Claude Desktop config file:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%\Claude\claude_desktop_config.json
  2. Paste the following JSON block into the mcpServers section:
{
  "mcpServers": {
    "rulip": {
      "type": "sse",
      "url": "https://mcp-trial-m3g4.onrender.com/sse",
      "headers": {
        "Authorization": "Bearer YOUR_PERSONAL_ACCESS_TOKEN"
      }
    }
  }
}

Note: Replace YOUR_PERSONAL_ACCESS_TOKEN with the token generated in your Rulip Developer Dashboard.

Restart Claude Desktop, and Claude will automatically be able to save and retrieve memories.

Cursor Cursor IDE

Connect Rulip to Cursor's built-in Composer or Chat panel to remember your project decisions across sessions.

  1. Open Cursor and navigate to Cursor Settings > Features > MCP.
  2. Click + Add New MCP Server.
  3. Configure the fields with these parameters:
    • Name: Rulip
    • Type: SSE
    • URL: https://mcp-trial-m3g4.onrender.com/sse?token=YOUR_PERSONAL_ACCESS_TOKEN
  4. Click Save.

Replace YOUR_PERSONAL_ACCESS_TOKEN with the Bearer token from your Rulip Dashboard.

Windsurf Windsurf IDE

Sync Rulip's persistent memory with the Windsurf IDE to let your AI coder access context from your web chat transcripts.

  1. Click the MCP button in the bottom status bar of Windsurf.
  2. Click the Settings icon to open the configuration editor.
  3. Add the following block under the mcpServers section:
{
  "mcpServers": {
    "rulip": {
      "type": "sse",
      "url": "https://mcp-trial-m3g4.onrender.com/sse",
      "headers": {
        "Authorization": "Bearer YOUR_PERSONAL_ACCESS_TOKEN"
      }
    }
  }
}

ChatGPT ChatGPT Custom Actions (OAuth)

For ChatGPT, Rulip supports full secure OAuth 2.0 authorization, letting you authorize GPTs in the browser without manually pasting keys.

  1. Create a Custom GPT in ChatGPT.
  2. Go to the Actions configuration section and add a new action.
  3. Import the OpenAPI schema from: https://mcp-trial-m3g4.onrender.com/static/openapi.json (or construct it from Rulip tools).
  4. Set the authentication type to OAuth with the following parameters:
    • Client ID: memorie-client
    • Client Secret: secret-123
    • Authorization URL: https://mcp-trial-m3g4.onrender.com/authorize
    • Token URL: https://mcp-trial-m3g4.onrender.com/token
  5. Save your Action and prompt the GPT to authorize.

one memory for all