Zendesk for StartupsUp to 2 years free for startups + AI Agents & Copilot
Promoted
Launched this week
chat-recall
Ctrl+F for every conversation you've had with an AI
64 followers
Ctrl+F for every conversation you've had with an AI
64 followers
Your team has done months of work with AI assistants. Claude Code, Codex, Cursor and OpenCode each keep a full record of it, in its own format, and none can read the others. One command reads what they wrote and turns it into one searchable history. Passwords come out before anything leaves your computer. Your assistant searches it itself, so it stops asking what you decided last month. It also finds keys that leaked into old chats, and checks which still work.
chat-recall
Hey Product Hunt 👋
Ctrl+F doesn't work on your brain. That was the whole problem I had.
I run four assistants — Claude Code, Codex, Cursor and OpenCode — depending on the job. Each one keeps a complete record of everything it did. Each one is blind to the other three. So I kept re-explaining the same architecture, and twice I watched one rebuild something another had finished the week before.
chat-recall reads what they already wrote and turns it into one searchable history. One command.
The part I did not plan for is the leaked keys. Old chats are full of pasted credentials, and it now tells you which ones still work. A dead key is cleanup. A live one is an incident.
Three things it does that I use every day:
- A new laptop already knows everything. Sign in and the whole history is there.
- Try a new assistant without starting over. Every add-on follows you across.
- Bugs turn into tasks on their own, with the fix already sketched out.
- Scan for passwords/secrets before sending any data.
Free forever for one person, on your own hardware. $10 a month if we host it. The whole repo is source-available under Elastic License 2.0.
Which assistant would you want indexed that I'm missing? I am happy to add more!
looks really cool and really OCD-friendly so i am gonna definitely subscribe. I have a question though. How does CR handle conflicting or outdated context? Like if Claude, Codex and Cursor reach different conclusions, or an earlier decision is later reversed, how does it determine the current source of truth, preserve provenance and prevent stale assumptions from being passed to the next "guy"?
chat-recall
@charlie_delshire Good question — yes I did some iterations on this. So for example I had an older auth solution then I have changed it - so I think this fits your exact situation.
Every fact is dated and keeps the session it came from, and the timeline returns them in order — so "chose Keycloak" in March and "chose BetterAuth" in September both come back, dated, each with the conversation behind it. An agent that asks sees the sequence and which came last.
What it doesn't do yet is close the old one by itself when the switch is only inferred from a transcript. Record the decision deliberately and the previous fact gets an end date.
For my personal use I am inclined to like this because I can see what happened when - but I see a use case in which I can record the decision - and even add some sort of functionality in which decisions are marked for architecture in general . So Ai can say : hey, I need to do auth - retrieve decision list >> this is what I should do.
Let me know if this makes sense and if its something you would be interested to have as a feature! Also let me know if you have any other feedback on it.
Does it pick up both Claude and ChatGPT history, or only the chats it sees in the browser?
chat-recall
@tomveber it picks up local chats using codex/claude code/opencode - so it is mainly oriented towards the local tool, then it parses its contents. So not oriented towards your browser. I could try to see if that is an option but current setup is focusing on developer context and interactions.
Think about it like this: you and your colleagues are working on a project and everyone is trying to do their own part - and at some point eventually you will need to understand what they did in order to integrate with their solution - this offers that context layer + any kind of scans/learnings/improvements derived from any conversation that you or they had.
I hope this brings more context. Happy to explore any possible options here!
This is going to save me a lot of time when I try to look for chats about certain topics. Good luck!
chat-recall
@elene_kvitsiani yes that was the goal indeed! I hope it's useful! Let me know if you have any feedback!
Great work! I needed this because I don't remember anything
chat-recall
@rekhviashvili haha! same here. Please let me know if you have any feedback!
Product Hunt
chat-recall
@curiouskitty the main reason was the need to sync between devices - need that I have + the team plan in which I really believe collaboration will improve if you can see what your colleague did to fix that bug. Same for skill sharing and mcp tool sharing - I've used postgres for a better scaling really but for a single dev sqlite could have worked as well.
Please let me know if you've tried it and if you have any feedback!
Paste
Came for the search, stayed for the leaked key detection. When it finds a key that still works, does it flag it or actually help you rotate it? Also wondering if it can pull in regular ChatGPT or Claude app chats, or just the coding assistants.
chat-recall
@protsenkoalexandra it flags it and adds a task - so there is a tasks list based on the scans and results - and it adds automatic tasks that AI can automatically fix and report through MCP - that part I still need some feedback on as for me it works but it might not be as obvious yet.
Also once flagged - it can add directly in claude.md or in agents.md a note automatically to report any secret leaked and ask to rotate on the spot.
So the goal here is not just to have a flag but with mcp to get ai to also fix the issues found for you.
I can look into getting regular chatgpt or claude apps - if it's local then I should be able to fetch the data. Would you want this in a future release ? or just asking what it can do.