Every AI chat starts from zero.
Yours doesn't have to.
Telekin is a long-term memory your AIs share, and you own. Tell one AI something, and the others already know.
You: My bike-shop site runs on Astro, deployed to Cloudflare. Bookings go through Cal.com.
Claude: Got it. I've saved that to your memory.
You: Add a "book a repair" button.
Cursor: I'll link it to your Cal.com booking page, styled like the rest of your Astro site…
Set up in 15 minutes
Connect your AIs
Add Telekin to every AI app you use.
5 minutes
Let it interview you
Your AI asks about your projects, tools and habits, and writes them down.
10 minutes
Just use your AI
Every new chat, in every app, starts out knowing your world.
from then on
Universal. Telekin works with any AI app that supports MCP, the open standard nearly all of them use now:
- Claude
- ChatGPT
- Claude Code
- Cursor
- VS Code
- Windsurf
- Codex
- Gemini CLI
- Cline
…and whatever you switch to next.
Memory you can trust an AI to write
Drafts first. Your AI's changes stay in a draft until they're published, so half-finished notes never show up.
Every note has a history. Who wrote it, when, and the sources it came from.
Undo anything. Every change is kept, back to day one.
Stays tidy. Checks for stale, bloated and broken notes as it grows.
Questions
How is this different from ChatGPT's or Claude's memory?
That memory stays inside one app, and you can't really see or correct it. Telekin is shared by every AI you connect, kept as readable notes (one per project, tool or person), and every change can be undone.
How is it different from hooking up Obsidian?
If you only use desktop apps and are happy running a local server, that works. Telekin is different in three ways. It's hosted, so it works in claude.ai on the web and on your phone, not just on the computer where your vault lives. It keeps your AI's memory separate from your own notes, so an AI can write freely and you can review and undo everything it did. And it looks after itself, checking for broken links and stale or bloated notes as it grows.
Where does my data live, and who can see it?
On Telekin's server, run by Hypotheosis. It's encrypted on the way between your AI app and the server. Only you, with your key, and the server's operator can access your memory. It's plain text, and you can ask for a full copy, or for it to be deleted, at any time. More about your data.
What will it cost?
It's free during early access. Plans will be announced before that ends.
Can I host it myself?
Yes. Telekin runs on okf-mcp, which is open source (AGPL-3.0) with every feature included.