Your memories are in ChatGPT... But nowhere else. Universal Memory MCP makes your memories available to every single LLM. No logins or paywall. One command to set it up.
Hi Product Hunt 👋
I'm Dhravya, and I have a problem. Over the last few years, I have been using ChatGPT a lot - and now that it has memory, it's great! But, Claude does not have this memory. Neither does Google, Windsurf, or Cursor.
So I made the Universal Memory MCP. With the Model Context Protocol becoming widely accepted, this enables you to connect your 'second brain' to every LLM and maintain context across them.
This is magical. I can discuss my project details with Claude and go to Windsurf and just say "Code the project we were talking about". And that's it. It works beautifully.
Been using this internally for a while. This is one of the BEST memory plugins + it's all open source and optimized. Dhravya is great at DX and customer support and ships fast.
🎉 Congratulations on launching Universal Memory MCP on Product Hunt! 🚀 It's exciting to see a tool that bridges the gap between different LLMs and makes our memories accessible everywhere. The idea of having a "second brain" that works seamlessly across platforms is truly innovative!
I love how easy it is to set up with just one command—no logins or paywalls!
🙌 One suggestion I have is to consider adding a feature that allows users to categorize or tag their memories for easier retrieval. This could enhance the user experience even further!
Looking forward to seeing how Universal Memory MCP evolves! Keep up the great work! 💪
Hola, me parece una idea genial, me podrias explciar mejor como funciona? yo creee en chatgpt un "comando" en su memoria para q base su tono, informacion, personalidad y demas cosas en tres archvios q cargo todas las mañanas y guardo al final del dia los mismos archivos q el mismo gpt me actualiza y guarda todo lo q se hizo en el dia. Lo estoy usando para crear un grupo de expertos q me ayuden a posicionar mi empresa en google, crear publicaciones para ig, textos adecuados para cada etapa (publicacion en tienda, textos persuasivos para la venta, a/b test para ir viendo q funciona mejor, ademas de poder recolectar informacion para generar publicos acordes a cada producto, etc) asi y todo, hay veces q se marea y devuelve info q no es correcta, lo cual no es grave si se advierte a tiempo, pero estoy en proceso de automatizar procesos con n8n y si hay algun error q no "distingo" puede crearse uan bola de nieve. q les parece mi sistema? lo bueno es q tambien puedo crear protocolos para usarlo con gemini, deepseek, etc, pero chat gpt es el q mas me gusta por el momento
Hey @dhravya I have been waiting for something like this! Awesome!
Quick querstion, the instructions and video give some clients such as claude ad witsy - can I change this part of the command: "....--client witsy" to say, chatgpt or gemini etc, and do I need to do that for each LLM I use? Thanks!
Replies
Supermemory
ConstitutionDAO (📜,📜)
It’s pretty impressive how easy this was to set up - no sign up form. just one line in terminal and it was installed.
Excited to see increased compatibility across various applications.
Sidekic AI
Dhravya and supermemory are exceptional. Excited for this launch.
Congrats on the launch. Trying this asap.
Been using this internally for a while. This is one of the BEST memory plugins + it's all open source and optimized. Dhravya is great at DX and customer support and ships fast.
how long does it take to sync everything up?
Supermemory
@miklesh_pal a few seconds at max!
Tate-A-Tate
🎉 Congratulations on launching Universal Memory MCP on Product Hunt! 🚀 It's exciting to see a tool that bridges the gap between different LLMs and makes our memories accessible everywhere. The idea of having a "second brain" that works seamlessly across platforms is truly innovative!
I love how easy it is to set up with just one command—no logins or paywalls!
🙌 One suggestion I have is to consider adding a feature that allows users to categorize or tag their memories for easier retrieval. This could enhance the user experience even further!
Looking forward to seeing how Universal Memory MCP evolves! Keep up the great work! 💪
Hatch Messenger
I love this concept. I trust hat GPT, probably too much. What if one of the linked up ai alternatives had a data breach or was unethical?
Supermemory
@dscornell You have to explicitly allow the MCP client to 'search' for every single query that it uses search. See the demo video for more details :)
Shopmatey
Congratulation on the launch!
I like the cross-LLM memory sync solution, super useful! 👍
@dhravya is doing some amazing work iterating in public. Memory is much needed for an agentic world! Go get em @dhravya
How secure is this? Is this encrypted provably?
Hola, me parece una idea genial, me podrias explciar mejor como funciona? yo creee en chatgpt un "comando" en su memoria para q base su tono, informacion, personalidad y demas cosas en tres archvios q cargo todas las mañanas y guardo al final del dia los mismos archivos q el mismo gpt me actualiza y guarda todo lo q se hizo en el dia. Lo estoy usando para crear un grupo de expertos q me ayuden a posicionar mi empresa en google, crear publicaciones para ig, textos adecuados para cada etapa (publicacion en tienda, textos persuasivos para la venta, a/b test para ir viendo q funciona mejor, ademas de poder recolectar informacion para generar publicos acordes a cada producto, etc) asi y todo, hay veces q se marea y devuelve info q no es correcta, lo cual no es grave si se advierte a tiempo, pero estoy en proceso de automatizar procesos con n8n y si hay algun error q no "distingo" puede crearse uan bola de nieve. q les parece mi sistema? lo bueno es q tambien puedo crear protocolos para usarlo con gemini, deepseek, etc, pero chat gpt es el q mas me gusta por el momento
Thanks for a project idea :).
So dope!
Hey @dhravya I have been waiting for something like this! Awesome!
Quick querstion, the instructions and video give some clients such as claude ad witsy - can I change this part of the command: "....--client witsy" to say, chatgpt or gemini etc, and do I need to do that for each LLM I use? Thanks!
I love how easy it is to get started. Congratulations on the launch!
Oooooo this is what we needed!