FAQ
Concise answers to the questions we get most often.
Is the agent autonomous or under my control?
Both. Hermes plans and acts on its own, but every plan step is visible and editable. You can pause, intervene, or abort at any time.
Does the agent ever sign on my behalf?
No. Hermes prepares transactions. You sign them in your wallet. There is no automatic-signing mode.
What if my local model can't do tool calls?
Hermes will fall back to a JSON-mode prompt that emits a tool call as structured text and parses it deterministically. Quality is somewhat worse but the loop still works.
How does CLAW NOVA compare to LangChain / AutoGPT / etc.?
LangChain is a library; CLAW NOVA is a complete workspace. AutoGPT was a research toy; Hermes is production-shaped (bounded, inspectable, interruptible). The workspace is the differentiator.
Can I use Anthropic / OpenAI / Together models?
Yes. Set your provider's base URL and key in Settings. Any OpenAI-compatible API works.
Is there a mobile version?
Not yet. The Hermes runtime is portable; a Tauri/native shell is on the roadmap.