Hermes Agent integration

Run Hermes Agent on Modellane models with a custom endpoint in its config file and the API key in its environment file.

Hermes Agent works with any OpenAI-compatible endpoint. Set Modellane as a custom provider in its config file and keep the key in its environment file.

Hermes Agent is an open-source agent from Nous Research. It works with any OpenAI-compatible endpoint; this page uses the Chat Completions API (POST /v1/chat/completions).

Configure the endpoint#

Put the key in ~/.hermes/.env:

Shell

MODELLANE_API_KEY=your-api-key

Then set the model in ~/.hermes/config.yaml:

YAML

model:
  default: lane-1
  provider: custom
  base_url: https://usemodellane.com/v1
  key_env: MODELLANE_API_KEY
  # Replace with the context length of the model from Models & Pricing.
  context_length: 128000
ParamValue
defaultlane-1, or any id from Models & pricing
base_urlhttps://usemodellane.com/v1
key_envThe variable in ~/.hermes/.env that holds your key
context_lengthThe context length of the model, in case Hermes cannot detect it

You can also run the hermes model wizard and choose a custom endpoint; it saves the same settings. Hermes supports several transports for providers. If you declare Modellane under providers instead, set transport: chat_completions.

Verify#

Start Hermes and send "Reply with the word ready." A reply means the endpoint works. An authentication error points at the key, and model_not_found at the model id.

Images#