Installation

Run CLAW NOVA

git clone https://github.com/clawnova/claw-nova
cd claw-nova
pnpm install
pnpm dev

Open http://localhost:3000.

Install a local model

We recommend Ollama for the default experience. Hermes will work with any OpenAI-compatible endpoint, but Ollama is the simplest path.

macOS / Linux

brew install ollama
ollama serve
ollama pull qwen2.5:7b

Windows

Download from https://ollama.com and launch the app. Then:

ollama pull qwen2.5:7b

Other backends

• vLLM — python -m vllm.entrypoints.openai.api_server --model Qwen/Qwen2.5-7B-Instruct
• llama.cpp — ./llama-server -m model.gguf --port 11434 --host 0.0.0.0 -c 16384
• LM Studio — enable the OpenAI-compatible server in Developer settings.

Configure CLAW NOVA

Open Settings → Model Provider and set:

• Base URL — e.g. http://localhost:11434/v1
• Model — e.g. qwen2.5:7b
• API key — leave blank for Ollama

Done. Hermes is ready.

Recommended models