Skip to main content
The Model Context Protocol (MCP) endpoint lets an AI application work with the same workspace resources you use in the dashboard, subject to your plan, role, and approved permissions.
On a white-label domain, use that domain instead. The login and consent screens use the same branding. Your plan must include Connect AI / MCP.

​
Connecting clients

  1. Open Settings → Connectors → Add custom connector.
  2. Enter https://app.famulor.io/mcp.
  3. Sign in and approve the requested permissions.
  4. The approved tools appear in Claude.

​
Connect from inside Famulor

Choose Connect in the sidebar, or Use with ChatGPT & Claude on the dashboard. The dialog shows the correct MCP URL for your current domain and offers starter prompts for common tasks — build an assistant, diagnose one that misbehaves, analyze your last call, improve from your last 30 calls, or launch a campaign.
Working from a coding agent such as Claude Code, Cursor, or Codex, install the open-source Famulor Skill with npx skills add bekservice/Famulor-Skill. It carries the endpoint configuration (add the URL manually if your client doesn’t pick it up) plus Famulor-specific working rules on top of raw MCP access: read current state before changing it, confirm anything hard to reverse, and stay inside the authenticated workspace.
Compatible clients start the OAuth sign-in automatically. The consent screen shows the requested permissions before anything is shared. Approvals can be revoked from your account settings. For a manual or server-to-server connection, send a workspace API key:
Use the narrowest scopes required. Write scopes include the corresponding read access.

​
Choosing which tools to expose

The standard endpoint exposes all tools allowed by the credential. For a smaller tool list, select one or more groups when connecting:
You can also send X-MCP-Toolsets: assistants,calls. The header takes precedence over the query parameter. Use list_mcp_toolsets to see the groups available to your account. The MCP client receives the current input schema, permissions, and safety hints for every tool it is allowed to use. This live discovery is the authoritative reference for individual tool arguments.

​
Public directory profile

The reviewed ChatGPT and Claude directory connection uses:
This profile is limited to read-only assistant information and read-only conversation history. It excludes assistant changes, outbound calls, campaigns, messaging, phone-number management, billing, integrations, and tool-group discovery. Custom MCP connections keep the permissions selected during their own setup.

​
Example prompts

Once a client is connected, plain language reaches every tool it’s approved for:
A well-behaved client reads current state before changing it, and asks you to confirm anything hard to reverse — an outbound call, a message send, a campaign start, a purchase. Use E.164 phone numbers in your prompts, review a generated system prompt before asking the client to save it, and keep one workflow per prompt so a multi-step request stays easy to follow.

​
Long-running tasks

Website imports, simulations, history exports, and large campaign preparations may continue as MCP tasks. A compatible client can show progress, retrieve the result, or request cancellation. Cancellation stops remaining work when possible; an action already accepted by an external service may still complete. Exports use short-lived private download links. Tasks and their results remain limited to the workspace and credential that created them. Call-history tools return a provider-neutral failure object when a call cannot complete. Use its code, customer-safe message, retry guidance, and suggested action instead of relying on infrastructure-specific details.

​
Errors

The same customer capabilities are available through the REST API. Choose the interface that best fits your integration.