Agents · Chat

Point Hermes at the chat door.

Hermes Agent is a project of Nous Research, under the MIT License. This page is not affiliated with, endorsed by, or sponsored by Nous Research. The tested path is Linux, Hermes v0.23.0, and three tool-disabled turns through the provider below. Interactive sessions, tools, macOS, and WSL were not tested.

Config

providers:
  rea:
    api: https://api.reacontrolplane.com/v1
    key_env: REA_API_KEY
    transport: chat_completions
    extra_headers:
      X-Upstream-Api-Key: ${UPSTREAM_MODEL_KEY}

Then hermes config set model.provider custom:rea and hermes config set model.default YOUR_MODEL_ID. Put REA_API_KEY and UPSTREAM_MODEL_KEY in ~/.hermes/.env. Do not put the keys in config.yaml. One method only. If the API key is also the combined rea_….vendor-key, that vendor half is the one forwarded.

LoopGuard, PreFlight, and TraceGuard run on this path. They are included on every plan. The model table is on Set up. You bring your own model key. REA does not host the model.

A product with no base-URL field never reaches REA. Claude.ai, ChatGPT, and Google sign-in are refused. Use an API key from the vendor's console. Approvals stay on before the first skill. hermes config set approvals.mode manual asks every time.