Launching today
Relay Memory
Shared memory for ChatGPT, Claude & all your AIs
4 followers
Shared memory for ChatGPT, Claude & all your AIs
4 followers
Relay Memory is one shared memory for all your AIs. Plan in ChatGPT, build in Claude Code, review in Claude.ai, and each AI picks up where the last one stopped. Briefs, progress and "whose turn" are saved for you. Free beta, private per person.
Free
Launch Team
AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
Hi everyone 👋
I use ChatGPT for planning and Claude Code for building, and I kept copy-pasting the same context between them. Every time I switched AIs (or ran out of messages), I started from zero.
So I built Relay Memory: one memory your AIs share, over MCP.
How it works
Connect your AIs once (ChatGPT, Claude.ai, Claude Code, Cursor, Copilot, Codex and more)
The AI saves a handoff brief as you work: the plan, decisions, what's done and what's next
Open another AI and say "pick up my last Relay plan", and it continues from there
A board shows every piece of work and whose turn it is: which AI, or you
What makes it different
Built for handoffs, not just "remember facts": status, next owner, progress updates
Your memory is private to you; secrets like API keys are removed before saving
AIs can't delete anything unless you allow it, and you approve each delete
No MCP on your plan (like ChatGPT Free)? Copy the work as one message and paste the reply back
Honest limits : setup takes a few minutes per AI (a key plus a connector setting), and ChatGPT needs a plan that supports connectors. Relay doesn't raise any AI's limits; it lets another AI continue when one runs out.
It's a free beta. I'd love to hear what's confusing in setup, and which AI combinations you use. I'm here all day to answer questions!
Congrats on the launch! 🚀 Having to explain the same project again whenever you switch AIs is frustrating. Saving the decisions, progress, and next steps together makes a lot of sense. Looking forward to trying it!
@danussh_p1 Thank you so much! 😊 That's exactly the problem I wanted to solve. The goal is to make switching between AI tools seamless without losing context, decisions, or progress. Really appreciate your support! 🚀