Tango MCP is in beta and available to Enterprise workspaces only. Access is limited to Enterprise workspaces that Tango has enabled. To request access for your workspace, reach out to your Tango customer success representative.
Tango MCP connects your AI assistant to the Tangos (step-by-step guides) your team already keeps in Tango. It uses the Model Context Protocol (MCP), an open standard for connecting AI tools to external data. Ask a how-to question in your assistant and Tango answers from the current guide, with a link to the source.
Before you start
Ensure that each of the following is true before you connect:
Your workspace is on an Enterprise plan and has been enabled for the Tango MCP beta.
You are a member of that workspace. Guests cannot connect.
Your AI assistant supports remote MCP servers with OAuth sign-in.
If your workspace doesn't appear when you sign in, it is either not on Enterprise, not yet enabled for the beta, or you are a Guest in it. Reach out to your customer success representative.
What you can do
Ask how to complete a task and get the matching Tango, with its steps and a link.
Ask follow-up questions about a specific step, including its screenshot.
Ask which Tango workspace you are connected to.
Connect your AI assistant
Using Microsoft 365 Copilot? It uses a separate install, which your Microsoft 365 admin completes. Reach out to your customer success representative for the Tango agent package. For all other AI assistants:
In your assistant's MCP or connector settings, add a remote MCP server using the details below.
When prompted, sign in to Tango and select the workspace to connect.
Test it by asking: "What Tango workspace am I connected to?" Tango replies with the workspace and account you selected.
Server URL: https://mcp.tango.us/mcp
Type: Remote MCP (Streamable HTTP)
Sign-in: OAuth 2.1
What results to expect
One Tango when there is a clear match. Shorter Tangos include the steps in the answer. For longer ones, open the link to see all steps.
Up to three Tangos when several may apply, so you can pick the right one.
No result when nothing suitable exists. Tango doesn't invent steps. Ask a Creator in your workspace to document the process.
Every result links to the current source Tango.
Tips for asking
Ask in plain language, the way you would ask a colleague. You don't need to know the Tango's title.
Name the task and the app, for example: "How do I submit an expense in Concur?"
Troubleshooting
I can't sign in or my workspace isn't listed. Check that your workspace is on Enterprise and enabled for the beta, and that you are a member rather than a Guest. Sign in with the Tango account that belongs to that workspace.
I get no results. Tango MCP only returns Tangos in the Team Library of the connected workspace that you have permission to view. Ask about a Tango you know is in that Team Library, and confirm which workspace you are connected to.
I need a different workspace. Each connection reads from one workspace. Disconnect Tango in your AI assistant, connect again, and pick the other workspace.
I can't connect at all. Confirm your AI assistant supports remote MCP servers and OAuth sign-in. If it does, reach out to your customer success representative.
Disconnecting
Disconnect any time from your AI assistant's connector or agent settings. There is no separate place in Tango to manage MCP connections. To change workspace, disconnect and connect again.
Security and access
Each person signs in with their own Tango account, and results follow that person's existing Tango permissions.
Access is read-only. The assistant can search and read Tangos. It can't create, edit, or delete anything in Tango.
Your Tango password never passes through the AI assistant.
Tango doesn't keep a separate copy of your content for AI. Every answer is read live from Tango at question time.
If you filter outbound traffic, allow
mcp.tango.us,app.tango.us,screenshots.tango.us, andimages.tango.us.
Beta limits
During the beta, Tango MCP can't create or edit Tangos, change permissions, support Guests or domain-wide access, or provide centralized administration or audit history across AI clients. Features may change as we learn from beta usage.
Questions and feedback
For account and access questions, reach out to your Tango customer success representative. For general questions about Tango MCP, to report a bad or missing result, or to share a feature idea, email [email protected].