MCP & API · 5 min · reviewed 2026-10-07
How do I connect my AI to more than one workspace?
Each workspace is its own CRM with its own address and its own tokens. A token from one workspace never opens another. To work in two workspaces, add one connector per workspace, each with its own address and token.
What you will need
- An account in each workspace you want to reach.
- An AI client that speaks MCP, such as Claude, ChatGPT or Cursor.
Steps
- 1
Step 1
Sign in to the first workspace. Go to My AI › Setup and create a token. Copy the address shown there. For SalesCrew it is https://<your-workspace>/functions/v1/salesops-mcp.
- 2
Step 2
In your AI client, add a server with that address and token. Give it a clear name, such as crm-clientA.
- 3
Step 3
Sign in to the next workspace. Create a token there too. Its address is different.
- 4
Step 4
Add a second server in your AI client with the second address and token. Name it crm-clientB.
- 5
Step 5
When you ask for something, name the workspace. For example: *"In crm-clientB, list today's tasks."* If your client signs in with a browser instead of a token, add the connector once per workspace address. Each one asks you to sign in to that workspace.
Common errors
- The AI shows the wrong client's data
- it used the other connector. Name the workspace in your request.
- 401 on the second workspace
- you used the first workspace's token. Each token works on its own workspace only.
- Tools are missing
- a module is off in that workspace, or your role there is narrower.
How you know it worked
Ask each connector *"which workspace am I in?"*. Each one names its own workspace. Settings › API access in each workspace lists only its own tokens.
Where this goes in SalesCrew
Each connector answers with data from its own workspace only. In your own workspace the same how-to opens beside the field, with your callback URLs filled in.