Connecting Rudow to your AI tools via MCP

1. Create an MCP key

1. Go to Settings → Account in the Rudow dashboard.
2. Under MCP keys, click Create MCP key.
3. Give it a name (e.g. Claude Desktop, work laptop).
4. Choose which projects it can access — leave All projects checked to let it follow your access as it changes, or uncheck it and pick specific projects to restrict the key.
5. Click Create key and copy it immediately — it's shown once, starts with alogmcp_, and can't be retrieved again (you'd need to create a new one).

Keys can be renamed or revoked any time from the same panel.

2. Connect a client

All clients need the same two things:

• Server URL: https://mcp.rudow.ai/mcp (Streamable HTTP transport)
• Auth header: Authorization: Bearer <your-alogmcp_-key> (or X-Api-Key: <your-alogmcp_-key>)

Claude Code

claude mcp add --transport http rudow https://mcp.rudow.ai/mcp \
  --header "Authorization: Bearer <your-alogmcp_-key>"

Or add it directly to .mcp.json:

{
  "mcpServers": {
    "rudow": {
      "type": "http",
      "url": "https://mcp.rudow.ai/mcp",
      "headers": { "Authorization": "Bearer <your-alogmcp_-key>" }
    }
  }
}

Claude Desktop / claude.ai

Go to Settings → Connectors → Add custom connector, paste in https://mcp.rudow.ai/mcp, and add an Authorization header with value Bearer <your-alogmcp_-key> when prompted for authentication.

Cursor

Add to .cursor/mcp.json (project-level) or your global MCP config:

{
  "mcpServers": {
    "rudow": {
      "url": "https://mcp.rudow.ai/mcp",
      "headers": { "Authorization": "Bearer <your-alogmcp_-key>" }
    }
  }
}

Windsurf

Add to your Windsurf MCP config (~/.codeium/windsurf/mcp_config.json):

{
  "mcpServers": {
    "rudow": {
      "serverUrl": "https://mcp.rudow.ai/mcp",
      "headers": { "Authorization": "Bearer <your-alogmcp_-key>" }
    }
  }
}

Codex CLI (OpenAI)

Add the server from the terminal:

codex mcp add rudow \
  --transport http \
  --url https://mcp.rudow.ai/mcp \
  --header "Authorization: Bearer <your-alogmcp_-key>"

Or add it directly to ~/.codex/config.toml:

[mcp_servers.rudow]
url = "https://mcp.rudow.ai/mcp"

[mcp_servers.rudow.http_headers]
Authorization = "Bearer <your-alogmcp_-key>"

OpenAI API (Responses / Agents)

The OpenAI Responses API can call remote MCP servers directly — pass Rudow as an mcp tool and the model gets the same log and metric tools as every other client.

const response = await openai.responses.create({
  model: "gpt-5",
  tools: [
    {
      type: "mcp",
      server_label: "rudow",
      server_url: "https://mcp.rudow.ai/mcp",
      headers: { Authorization: "Bearer <your-alogmcp_-key>" },
      require_approval: "never",
    },
  ],
  input: "List my Rudow projects and show the error rate for the last hour.",
});

Any other MCP client

If your tool supports remote MCP servers over Streamable HTTP, use the same URL and header — the config field names above (url/headers or serverUrl) are the most common shape, but check your client's docs if it uses different field names.

3. What you can ask it to do

Once connected, the assistant has access to these tools:

Every tool (other than the two discovery ones) is scoped to a project_id and checked against your key's project access on every call — a key can never read a project it isn't allowed to.

4. Example prompts to try

• “List my projects, then show me the error rate for the last hour on each.”
• “Search my logs for anything mentioning 'timeout' in the last 30 minutes.”
• “What are my top 5 slowest routes right now, and what's driving the latency?”
• “Something's wrong with checkout — dig into the logs and tell me what's happening.”
• “Help me add Rudow logging to this Node service.” (uses list_sdks to fetch setup instructions)

5. Usage limits

MCP requests are metered per month based on your plan (see Settings → Billing → MCP requests this month). Discovery calls (list_projects, list_sdks) don't count against this limit — only calls that actually query your data do.

6. Troubleshooting