Hermes Agent Model Provider and API Key Setup

Hermes Agent works with any OpenAI-compatible API: run hermes model to pick a provider and set your key, or choose Custom endpoint for local models like Ollama.

Who this is for：Anyone freshly installed and stuck on which provider to pick or how to set the API key.

Steps

  1. Open the provider menu

    Officially supported: OpenRouter, OpenAI, Anthropic Claude, Google Gemini, and local models via Ollama/vLLM.

    hermes model

  2. Or use the official Portal

    Access 300+ models through Nous Portal.

    hermes setup --portal

  3. Set a key directly (advanced)

    hermes config set OPENROUTER_API_KEY your_key

  4. Local models (Ollama example)

    Run hermes model, choose Custom endpoint, enter your local address, and set context length to your server's real capability.

    http://localhost:11434/v1

  5. Pin a specific model

    hermes config set HERMES_MODEL anthropic/claude-opus-4.7

How to verify

FAQ

API key set but still failing?

Official FAQ: check with hermes config show, then reconfigure via hermes model. Keys must match the provider: an OpenAI key does not work with OpenRouter.

How much does Hermes cost?

Hermes Agent itself is free and open-source (MIT). You only pay your provider's LLM API usage; community reports show provider choice can change spend by an order of magnitude.

Cheaper model for subagents?

Yes: configure the delegation block in config.yaml, e.g. model: google/gemini-3-flash-preview with provider: openrouter.

Context length exceeded?

Compress the session with /compress, check /usage, or set model.context_length explicitly in config.yaml (e.g. 131072).

Related pages