You can now run @supermemory locally. Introducing the supermemory local - Fully self-contained. Comes with our graph engine, embedding model, etc. - Run on any machine, with your @openclaw, hermes, claude, etc. - SDKs to add memory to your agent, or build your company brain.

tomorrow.

we’re donating $10 to charity for every person that shares a memory with me! see you at Dolores :)

I made a universal company brain. - connect to ANY agent natively - Git-like versioning and RBAC permissioning - connectors to all sync tools of companies - Dreams about your company - run on-prem. Free to start. it's live today. we've been using it for 6 months @supermemory

We set out to find out how we can build the best filesystem for agents out there. SMFS is consistently better in terms of tokens, cost, latency AND accuracy. Technical report here: smfs.ai/research/memory-as-a…

Agent observability can be more than logs and traces. Introducing Qualitative Analysis. it's FREE for next 3 months 👇 @Supermemory can now automatically give you reports on your users/agents - which you can tailor and ask questions to! Eg: "What's the split of developers?"

founders after @supermemory

The Context Cloud for agents is here. Introducing the new @supermemory.  A new way to do context, for your agents Everything you need for context - Memory, RAG, Filesystems, Profiles, are now available as blocks you can compose to build something for your use case

Agent can grep, find, cat and append using new preprint representation

Today, we're switching to Dynamic Dreaming in the supermemory API It's an interesting and honestly, magical new way to do memory. I wrote a bit about it here 👇

wtf. Web pages as text files?!?! Instead of vision and images, Preprint enables LLMs to edit the markdown files to do actions on the browser. This is insane!

What if project the web Into the substrate LLMs already work with github.com/supermemoryai/pre…

Fun story: Kush helped me build the v0 of supermemory, 2 years ago. We kept in touch, and now he's in the team, handling the infrastructure now that usage is exploding!

We built an AI hackathon judge! Agentic evaluators are a fascinating, fast-moving space. We ran an experiment for a recent hackathon in partnership with @AnthropicAI and @genspark_ai, and learned a ton about where agents help, where they fail, and where humans need to stay in the loop. Wrote about those lessons, and how we're thinking about making project and builder evaluation fairer, more evidence-based, and more robust at scale 👇

Life Update: Joined @supermemory to make your agents unforgettable!