15 points boulevard | 1 comments | 23 Aug 25 04:57 UTC | HN request time: 0.199s
| source AgentState to solve a problem I kept running into: managing state for multi-agent AI systems is surprisingly hard.
When you have multiple AI agents that need to coordinate, persist their state, and query each other's status, you typically end up with a mess of Redis/Postgres setups, custom queuing, and manual synchronization code.
The whole thing is ~3MB, written in Rust for performance and safety, runs in Docker, and handles 1000+ ops/sec. I've been running it in production for AI workflows and it's been rock solid.