Your Chat.co experts in your AI apps

1. How it works

Chat.co runs one connector server at https://mcp.chat.co/mcp, using the Model Context Protocol (MCP) that many AI apps support. The app connects to it, you approve the connection in Chat.co, and from then on the app can ask the experts you already use in Chat.co.

2. Choose your AI app

Each app has its own setup guide. Availability differs: a directory listing is only described as published once the vendor has published it, and a status we could not confirm is described as unconfirmed. Status reviewed 30 September 2026.

3. Connecting your account

Whichever app you use, you add Chat.co from that app's connection settings and sign in to your own Chat.co account. You never need an API key or another person's login.

Every host sends you to the same Chat.co approval page. Sign in if asked, then check each part before you approve:

Choose Approve connection to finish, or Cancel to stop. The only permission requested is chatco:read. It means the app cannot change your chatbots, knowledge, settings or account. It does not make the connection free of side effects: asking an expert saves the conversation in Chat.co and uses message credits, and a standalone search uses credits too, exactly as in the Chat.co app. Your questions, and the answers, excerpts and files you are permitted to receive, pass to the AI app and its provider.

4. Experts and sources

If you can chat with an expert in Chat.co, you can ask that expert through the connector. Follow-up questions stay in the same expert conversation. Your current account, workspace and licence permissions apply to every request.

Document-based answers include citations and links to the supporting documents. You don't need permission to manage an expert's knowledge to open the evidence for an answer you were allowed to receive. Private documents require sign-in and current access; deleted or unavailable sources may no longer open.

Searching documents directly is a separate capability. It is available only when your account may retrieve that expert's knowledge, and the expert list shows whether you can.

5. Message credits

A search that completes with no matches still counts. Failed operations, inaccessible experts and retries of the same operation are not charged again. Your existing allocation applies; there is no checkout inside the connector. When credits run out, new paid requests stop, but you can still list experts and open sources you already received.

Manage your allocation through your usual Chat.co account process.

6. Data and privacy

A connected app can send your questions to Chat.co and receive the authorized answers, source excerpts, citations and document links, plus any assessment or file you ask an expert to produce. Where requested during sign-in, it also receives your account identifier, email address and email-verification status. It cannot manage or delete your chatbots through this connector.

The app's own terms and privacy policy govern the information it receives. Disconnecting stops future access but cannot remove content already delivered to that app. Chat.co's Privacy Policy and Terms apply to your use of Chat.co.

7. Manage your connections

Open Connected applications to review each app and its workspace, or disconnect it. To reconnect, start again from the app's connection settings.

If an expert is missing, check that you chose the right workspace and still have access in Chat.co. If a private document doesn't open, sign in with the account that received the answer. Each setup guide lists the errors specific to its app.

8. For developers & agents

You don't need this section to use Chat.co from an AI app. It documents the protocol for developers and for agents that call the tools directly.

Server

Tools

* Required.

Recipe for agents

1. 1

   List experts first

   Call list_authorized_experts. Only use id values it returns. Offer search only for experts whose access.knowledgeSearch is true.

2. 2

   Consult with a fresh operation_id

   Generate a new identifier (a UUID works) for each new question. Keep the returned conversation_id.

   consult_expert argumentsjson

   Copy

   {
     "expert_id": "<id from list_authorized_experts>",
     "question": "How many vacation days does a new employee get? Cite the sources.",
     "operation_id": "7f3b2c1e-5a9d-4e6f-8b0a-2d4c6e8f1a3b"
   }

   result (abridged)json

   Copy

   {
     "conversation_id": "<keep for follow-ups>",
     "answer": "New employees receive … [1]",
     "citations": [
       {
         "id": "<source id>",
         "title": "Employee handbook",
         "url": "<source link>"
       }
     ]
   }

3. 3

   Follow up in the same conversation

   Pass the conversation_id back with a new operation_id.

   consult_expert argumentsjson

   Copy

   {
     "expert_id": "<same expert>",
     "conversation_id": "<conversation_id from the first answer>",
     "question": "Does that change after two years of service?",
     "operation_id": "c41e9d27-0b6a-4f38-9e15-7a2d8c3b6f90"
   }

4. 4

   Read evidence with fetch

   Pass the citation's id and its title. The title is shown in approval prompts; the id is the lookup key. The excerpt may not be the whole document.

   fetch argumentsjson

   Copy

   {
     "id": "<source id from the citation>",
     "title": "Employee handbook"
   }

5. 5

   Present the answer with its links

   Show every cited source as a link to its url. Offer entries under files as downloads, not as citations. Don't add credit rates, model names or internal configuration to the answer.

Synthetic example: the HR Handbook expert, its answer and the angle-bracket values are placeholders. Use the identifiers and source links the server actually returns.

Claude, ChatGPT, OpenAI, Codex, Cursor, Grok and xAI are trademarks of their respective owners. Their names and logos appear only to identify compatible apps and do not imply endorsement.