Quick Start
Get Rhesis running locally in under 5 minutes. This setup is for local testing only.
Local Development Only
This configuration is NOT suitable for production. It uses generated secrets and enables local authentication bypass (QUICK_START=true).
Default Access
Once running, access the platform at:
| Service | URL | Credentials |
|---|---|---|
| Frontend | http://localhost:3000 | Auto-login (no credentials needed) |
| Backend API | http://localhost:8080/docs | API documentation |
Connect a Model
Rhesis needs a language model to generate tests and evaluate results. Until one works, the app opens on a Connect a model step instead of the dashboard. Pick one of its two options:
- Use a Rhesis API key runs on the models hosted by Rhesis. Get a key at app.rhesis.ai/tokens and paste it in.
- Add your own model connects a provider account (OpenAI, Anthropic, Gemini, and others) or a local model (Ollama, vLLM, LiteLLM proxy). See Models for the connection fields.
The step closes as soon as the model works. No restart is needed.
./rh start also asks for a Rhesis API key while .env.docker.local has none. A key entered there is saved as RHESIS_API_KEY in .env.docker.local, and the app skips the step. To change that key later, edit the file and run ./rh restart.
Evaluating only with code metrics? A test run started from the SDK or API whose metrics are all SDK code metrics needs no evaluation model.
Optional Configuration
Managing Services
Use the ./rh CLI to manage services:
Changing the ports
The first ./rh start prompts for backend and frontend ports (defaults: 8080, 3000) and writes them to .env.docker.local. To change them later, edit .env.docker.local — the port and the port in its URL must stay in sync (BACKEND_PORT with API_BASE_URL, FRONTEND_PORT with FRONTEND_URL). Then run ./rh restart:
Next Steps
Ready for Production?
For production deployment, see the Docker Compose Guide.
For local development setup without Docker (for contributing to Rhesis), see the Environment Setup Guide.