Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.monaco.com/llms.txt

Use this file to discover all available pages before exploring further.

The Monaco MCP server connects AI agents like Claude, ChatGPT, Cowork, and custom MCP clients directly to your Monaco workspace. Agents can search, update, and act on your Monaco data using natural language, with no custom integration required.

Use the Monaco MCP to

  • Start outbound to a new segment without leaving your AI client. Find contacts matching firmographic and demographic criteria, tag them as an audience, and enroll them in a sequence – all in one conversation.
  • Get up to speed on an account before a meeting. “Show me everything on Acme Corp – the open opportunities, the contacts I’ve engaged with, our recent meetings, and what’s outstanding.” One prompt, one summary.
  • Keep your pipeline clean. Run pipeline hygiene in a single prompt – find opportunities with slipped close dates, deals stuck in the same stage too long, or accounts that have gone dark, then update or close them in bulk. Operations Monaco’s UI doesn’t expose for opportunities.
  • Build a custom revenue dashboard from your CRM data. Compute weekly cohort win rates, monthly net new ARR, weighted pipeline forecasts, and loss-reason breakdowns – directly from your opportunities, with the analytical depth a strong AI can compose itself.

Getting started

The Monaco MCP server is available at https://mcp.monaco.com/mcp. Authentication is OAuth 2.0. Your MCP client handles the flow on first connection.
If your AI client is part of a corporate or managed workspace, you may need administrator permission to add a custom MCP connector.

Claude Desktop / Cowork

  1. Open Settings → Connectors
  2. Click Add custom connector
  3. Enter the URL https://mcp.monaco.com/mcp
  4. Complete the Monaco login in the browser window that opens
  5. The Monaco tools appear in your tool list

ChatGPT

Custom MCP connectors in ChatGPT are available on Business, Enterprise, and Edu plans. They require Developer Mode to be enabled at the workspace level first, which is typically an admin action.
  1. As a workspace admin, open Workspace Settings → Permissions & Roles → Connected Data and enable Developer mode / Create custom MCP connectors.
  2. In ChatGPT, open Settings → Connectors → Create.
  3. Enter a name (e.g. “Monaco”), a short description, and the URL https://mcp.monaco.com/mcp.
  4. Complete the Monaco login when prompted.
  5. In any conversation, open the chat menu, select Connected apps, and enable the Monaco connector.

Connecting with API Key

You can also use API key auth with Codex and ChatGPT to connect to the MCP server. codex mcp add monaco --url https://mcp.stage.monaco.com/mcp --bearer-token-env-var MONACO_MCP_TOKEN

Custom MCP client

Any MCP-compliant client can connect to https://mcp.monaco.com/mcp using OAuth 2.0. Configure your client with the server URL and complete the authorization flow.

Supported tools

All tools enforce your existing Monaco user permissions. Reads are auto-approved by your AI client; writes and deletes prompt for confirmation by default.

Records and objects

ToolDescription
list_accounts, get_accountSearch and fetch accounts. Supports filters, sorting, and pagination.
list_contacts, get_contactSearch and fetch contacts.
list_opportunities, get_opportunitySearch and fetch opportunities.
list_meetings, get_meetingSearch and fetch meetings. Get returns the AI summary and full transcript.
list_tasks, get_taskSearch and fetch tasks.
create_contact, update_contact, delete_contactCreate, update, or delete contacts.
create_opportunity, update_opportunity, delete_opportunityCreate, update, or delete opportunities.
upsert_account, delete_accountUpsert or delete accounts.
create_task, update_taskCreate or update tasks.
list_sequences, get_sequenceList or get sequences. Get sequences returns tasks.
list_sequence_templates, get_sequence_templateList or get sequences templates. Get sequences templates returns tasks and transitions.

Sample prompts

Building a new outbound audience

“Find all contacts at companies in San Francisco with title CEO or Founder who aren’t already in any active sequence. Create a new contact tag called Q4 SF Founders, apply it to them, then enroll each one in the Founder Outreach sequence template.”
One conversation; the agent handles the filtering, tag creation, tag application, and enrollment.

Prepping for a customer call

“Pull up everything I have on Acme Corp – open opportunities, recent meetings with their AI summaries, the contacts I’ve engaged with, and any outstanding tasks. Summarize where the relationship stands.”
The agent stitches together accounts, contacts, opportunities, meetings, and tasks into a single briefing.

Running pipeline hygiene

“Find all opportunities in Proposal stage with estimated close dates more than 30 days in the past. Update them to Closed Lost and add a note: ‘No response after follow-up.’”

Building a custom revenue dashboard

“Build me a revenue dashboard from my opportunities. Show weekly cohort win rates, monthly net new ARR, my open pipeline broken down by stage with a weighted forecast (proposal weighted high, qualified medium, meeting booked low), and the top reasons we’re losing deals – pull those from the closed-lost notes.”
Because the MCP connects to live Monaco data, the dashboard stays current every time you re-run the prompt. We recommend using a code-capable client like Claude Code, Codex, or Cowork for this kind of workflow.

Permissions

The Monaco MCP server uses OAuth 2.0 for authentication. Every request is authenticated as the user who completed the login, and every tool call respects that user’s existing Monaco permissions.

Rate limits

The Monaco MCP server shares rate limits with the public REST API: 100 requests per minute per organization, sliding window. Limits apply across all MCP tool calls and direct REST calls regardless of which user is making the request. If you hit a rate limit, the server returns HTTP 429. Compatible MCP clients back off and retry automatically.