YourMemory
Your AI forgets you. YourMemory doesn't
Start new thread

YourMemory 2.0 - Your AI has the memory of a goldfish. Not anymore

by•
Most AI memory tools (Mem0, Zep, vector DBs) are storage with search, they hoard everything and bloat into noise. YourMemory behaves like a brain: 🧠 It consolidates related facts into clean summaries, so memory gets sharper, not bigger (444 → 16). ♻️ It recovers your agent's context after every reset. 🔒 It proves every change with a tamper-evident audit trail. 👥 It shares role-based team memory pools. Everyone else's store gets worse as it grows. Ours gets better.

Add a comment

Replies

Best
Maker
📌

Hey Product Hunt 👋 I'm Sachit, maker of YourMemory.

I built this because I was tired of my AI agent treating me like a stranger every single morning. Same context, re-explained. Same preferences, forgotten. Every session started from zero.

So I gave it memory. But raw memory hits the flaw every tool in this space runs into, it just hoards everything and slowly drowns in its own noise.

That's the problem this relaunch solves. I rebuilt YourMemory to work like a brain, not a database:

🧠 It consolidates, when enough related facts pile up, it compresses them into one clean summary and archives the rest. Memory gets sharper over time, not bloated. (On my own store: 444 memories → 16 summaries.)

📉 It forgets on purpose, memories age on an Ebbinghaus forgetting curve, so stale facts fade while the things you actually use stay strong. Forgetting the right things is a feature, not a bug.

♻️ It survives context resets, when the context window compacts, it hands the working context back, so the agent doesn't re-read everything to figure out where it was.

🔒 It's provable, every read/write/delete is logged in a hash-chained, tamper-evident audit trail. Plus export + one-click right-to-forget.

👥 It's shared, role-based team memory pools, so your whole team's agents draw on the same institutional knowledge.

MCP-native (Claude, Cursor, any client), open source, and self-hostable, your memories never leave your machine.

▶ Try the live interactive demo (30 sec, no signup):

I shipped ~19 releases in one week to get here, and I'd genuinely love your feedback. 🙏

One question for you: what's the one thing you wish your AI agent would never forget about you or your work? 👇

 love the idea, rooting for you, congratulations on the launch!

Plugged it into my Cursor setup yesterday and the context actually carried over the next morning without me re-explaining half my stack. The audit log is a nice touch for debugging weird agent behavior too.

the tamper-evident audit trail is a really thoughtful touch, love that memory changes are tracked so you can actually trust what's stored over time.

Maker

 Thank you, that's exactly the intent 🙏 Memory you can trust, not just store. Every read/write/delete is hash-chained, so if a single record is ever altered the chain breaks and verify catches it. On top of that there's export and one-click right-to-forget, which is what makes teams comfortable actually running it on real data.

Curious from your side, when you think about trusting an agent's memory over time, is it more about auditability (proving what changed) or control (being able to correct/forget)? Trying to figure out which one matters more to people. 👀

Plugged it into my Claude setup and the consolidation is genuinely useful, way less context bloat than my old notes dump. Recovery after a fresh session felt seamless.

The tamper-evident audit trail is a really thoughtful touch, especially for teams sharing memory pools where trust and traceability actually matter.