been running hermes agent by nous research for a bit now and the local model support is genuinely better than anything else ive tried in this space
the thing that sold me: it has per-model tool call parsers built in, so it actually handles tool calling properly on 30B class models where openclaw and most other frameworks just fall apart. multiple people on here have confirmed its way less token hungry too
the self improving skills thing is real but honcho (the learning engine) is off by default which confused me for like 2 days before I figured it out.. once you enable it in config.yaml the difference is noticeable within a few sessions
some stuff worth knowing:
one command install, handles python and node and everything. supports ollama, vllm, sglang out of the box. six terminal backends including modal and daytona for serverless (basically free when idle). gateway connects telegram discord slack whatsapp signal from one process. built in openclaw migration if youre switching
the team behind it (nous research) are the people who actually train the hermes model family so the integration between agent and model is tighter than anything a third party wrapper can do
22k stars, MIT license, no telemetry. v0.6.0 just dropped with profiles for multi-instance setups
wrote up a full deep dive with install paths, real costs, security breakdown, comparison vs openclaw/cowork/creao: here
curious what models people here are running with it.. ive been on mistral 30b via ollama and its surprisingly capable for the size
[link] [comments]