Hey Product Hunt, builder here. The frustration that started this: I use four AI tools daily and every single one starts from zero each session. I kept writing the same context paragraph over and over. OMP is my attempt to fix the layer underneath that problem. The spec is open, the server is Apache 2.0, and there is a live hosted version you can point any tool at right now with no setup. The bit people seem to find most useful is the handoff brief, when you open Claude after a ChatGPT session, the extension generates a natural paragraph summarising what you were working on and injects it automatically. The new AI picks up mid-thought. Would love feedback on the protocol design itself. If you build AI tools and want to add OMP support the adapter interface is small, happy to help anyone who wants to build one. Try the live server here: https://omp-server-production.up...

open-memory-protocol

Stop repeating yourself to every AI tool you use

Stop repeating yourself to every AI tool you use