Let an AI assistant read your notes without uploading them
You want the assistant to know what you wrote last Tuesday. The obvious route is to paste it in, or to keep everything in a service the model already reads. Both end with your notes on somebody else's machine.
What people usually do
- 1
Copy and paste into the conversation
Works for one note. It does not scale, and whatever you paste reaches the provider anyway.
- 2
Keep the notes in a service the model can reach
Convenient, and it means your notes live there from now on, under its terms and its outages.
- 3
Upload a folder as attachments
A snapshot, not a connection. Anything you write afterwards is not in it.
What that costs
- Your notes stop being only yours: they sit on a server, under terms you did not write.
- It goes one way. The assistant reads what you sent, but cannot write back into your notes.
- Every conversation starts from nothing unless you keep sending the same context again.
Opening a door instead of handing over the room
- The Model Context Protocol lets an assistant talk to a program on your own machine, with your permission, rather than to a copy of your data in the cloud.
- Tisty runs an MCP server, so the assistant reads and writes your notes, documents and tasks where they already are.
- The files never leave the machine: what travels is the answer to the question you asked.
- Free and open source, for Windows and macOS.
Carry on with
Questions
What is MCP?
The Model Context Protocol, an open standard that lets an assistant talk to programs and data sources. Instead of uploading your files, the assistant asks a program running on your machine, and that program decides what to answer.
Do my notes reach the model provider?
Only what the assistant reads in order to answer a given question travels, inside the conversation itself. The files stay on your disk; nothing is uploaded or synchronised in the background.