MemoryBridgeAI

One private memory, shared by ChatGPT, Claude, and Cursor

9 followers

One private memory, shared by ChatGPT, Claude, and Cursor

9 followers

Your AI tools each remember a different you, and none of them share notes. MemoryBridge is one private memory layer they all read: connect Claude.ai, ChatGPT, Cursor, and Claude Code in about two minutes via hosted MCP, or import your ChatGPT/Claude history to start with a full memory on day one. Every auto-captured memory waits for your approval, every chat answer cites its source, and PII is scrubbed before anything is stored. Your context, portable and auditable.

This is the 2nd launch from MemoryBridgeAI. View more

MemoryBridgeAI v2.0

Launching today

One private memory, shared by ChatGPT, Claude, and Cursor

Free Options

Launch tags:Productivity•Developer Tools•Artificial Intelligence

Launch Team / Built With

Framer AI AgentsDesign and publish professional sites with AI

Promoted

Maker

📌

Hey Product Hunt, I'm back. Ten months ago I launched MemoryBridge here as a rough alpha with a config file and a dream: your AI tools should remember you, not the other way around. Since then I have been building nights and weekends with feedback from our alpha users, and 2.0 is a different product. The problem hasn't changed. ChatGPT knows your coding style. Claude knows your writing voice. Cursor knows your project. None of them share notes, and every new chat starts with you re-introducing yourself. The big labs are each shipping their own memory, which makes it worse: now your context is fragmenting into more walled gardens, not fewer. What's new since v1: - Two-minute connections, no config. MemoryBridge is now a hosted MCP server with OAuth. Connect Claude.ai web, ChatGPT (Developer Mode), Cursor, and Claude Code by pasting one URL. The "requires some config" caveat from my first launch is gone. - Start with a full memory, not an empty one. Import your ChatGPT or Claude export, PDFs, or Markdown. We extract candidate memories automatically so day one feels like month six. - You approve what gets remembered. Auto-extracted memories land in a review Inbox first. Nothing becomes part of your memory without your sign-off. Your AI takes notes; you hold the pen. - Answers you can verify. Ask your memory anything and every answer cites the exact memories it used. Click the citation, see the source. - A real trust layer. PII (keys, cards, SSNs) is automatically scrubbed before storage. Each connected tool gets its own permission scopes, so your coding assistant can read work context without ever touching personal memories. Encrypted in transit and at rest, never used for training, export and delete anytime. - Memory that stays fresh. Retrieval favors recent, relevant context, and a health view shows you which tool read what, plus what has gone stale. [Insert 2-3 honest traction numbers here: memories stored across alpha users, retrievals served, connected clients. Real numbers, even small ones, beat adjectives.] A promise kept: everyone from the first-100 alpha keeps free lifetime access to the core tier, exactly as I said 10 months ago. For this community: [promo details]. Why me: I led the engineering team behind the ChatGPT web and mobile clients at OpenAI and now lead consumer product engineering at Reddit. I have watched context get trapped inside every one of these tools, and I am building the layer I always wanted: memory that belongs to the user. I would love your take on two things: which tool should we connect next, and what would make you trust an AI memory product with your real context? Let's build the future of AI memory together. The story so far: our September 2025 launch lives here — https://www.producthunt.com/prod... — thanks to everyone who supported v1.

Report

5d ago

Previous MemoryBridgeAI Launches

MemoryBridgeAIPersonal Memory Access Across All AI Assistants and Tools.

Launched on September 29th, 2025