How do I connect any AI assistant to my workspace?
The Cloud runs a Model Context Protocol (MCP) server. Any MCP-capable AI assistant — such as Claude, ChatGPT, or Cursor — can connect to it. Once you authorize the connection, that assistant can read your memory and act in your workspace within the permissions you grant. You start at thecloud.so/connect, approve the connection, and the assistant gains scoped, consented access.
3 min read
What MCP gives you
The Model Context Protocol is an open standard for letting AI assistants talk to external tools and data. The Cloud exposes your workspace over MCP, so an assistant is not limited to what you paste into a chat — it can read your memory and use your workspace's tools directly.
Because it is a standard, you are not locked to one AI. The same connection works for any assistant that speaks MCP.
Connecting an assistant
Open thecloud.so/connect and start the connection for the assistant you use. You will be asked to approve access — this is the consent step, and it is explicit.
Once approved, the assistant can recall what you have saved and take actions you have permitted. Nothing happens without that authorization, and you can revoke it at any time.
Consent is the floor
Access is private by default. An assistant only sees what you allow, every grant is scoped, and sharing or changing visibility is always a per-action choice you make.
This is what makes cross-AI memory safe to use: the memory is yours, the permissions are yours, and the assistant is a guest you invite in.
Common questions
Which AI assistants can connect to The Cloud?
Any assistant that supports the Model Context Protocol (MCP), including Claude, ChatGPT, and Cursor.
Can an AI change things in my workspace without asking?
No. Access is private by default and every connection is consented and scoped. You grant permissions explicitly and can revoke them at any time.
Where do I start the connection?
At thecloud.so/connect, where you authorize the assistant to connect.