Hi PH 👋 I'm Seacen, BU Digital Technology Head for Greater China at Unilever — so by day I'm on the front line of how a big company actually adopts AI, and I kept hitting the same wall: the genuinely capable agents can do the work, but they can't see *my* work. I built OMem in my own time to close that gap, and I've been living on it daily for a couple of months — it's how I hand my own projects' context to the agent I build with. The thing I most want to find out now is whether it gives other people that same "oh, the AI finally gets it" moment. Two things I cared about: it's local-first (your work context is plain Markdown on your own disk, not a vendor cloud — you can read it, grep it, take it anywhere), and it's agent-agnostic on purpose, because the hot agent changes every year but your work memory shouldn't be locked to any one of them. Happy to answer anything in the comments.

OMem

Local-first work memory for your AI agents

Local-first work memory for your AI agents