Hermes UI

A dynamic, web-based UI for the Hermes autonomous AI agent. This interface bridges the gap between terminal hackers and non-technical users, offering a beautiful, guided experience for configuring AI providers and interacting with the hermes CLI.

Key Features

• Guided Setup Wizard: Seamlessly detects your installation and walks you through configuring either Cloud (OpenRouter, Anthropic, etc.) or Local (Ollama, llama.cpp, LM Studio) AI engines.
• Messaging Integrations: Quickly check the status of and get connection instructions for Telegram, Discord, and WhatsApp bridges.
• Rich Chat Interface: Real-time streaming, syntax highlighting, session management, and granular agent tool activity monitoring.
• Under-the-hood Filtering: Automatically strips out AI boilerplate, disclaimers, and verbose <think> reasoning chunks for a cleaner read.

Prerequisites

• Python 3.10+
• Node.js
• Hermes CLI agent (Installed natively or via Docker)

Setup

1. Backend

The backend is a FastAPI server that manages WebSockets to communicate with the Hermes CLI.

cd backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn server:app --reload

2. Frontend

The frontend is a React application built with Vite.

# In a new terminal, from the root of the project:
npm install
npm run dev

3. Usage

Open http://localhost:5173 in your browser. The Setup Wizard will automatically appear if you haven't configured Hermes yet, allowing you to choose between a Cloud provider API key or connecting your local inference engine.

---

🐳 Docker (Production)

The Docker setup runs the backend and serves the built React frontend from a single container.

docker compose up -d --build

The UI will be available at http://localhost:8000.

Tip

Local Inference Compatibility: If you use the setup wizard to connect to Ollama or llama.cpp while Hermes is running in Docker, you should use host.docker.internal as your hostname instead of localhost. The wizard's "Test Connection" button will verify if the container can reach your host machine's LLM engine.

• Your ~/.hermes directory is mounted into the container so config changes and sessions persist across restarts.