Question 1

Is the Community Edition limited?

Accepted Answer

No. The workflow that domain experts and engineers actually use is in the open version: reviews, annotated test sets, behaviors, metrics, generation, simulation, tracing, experiments, the SDK, and MCP. The Enterprise Edition adds SSO, RBAC, and API Clients, plus support and onboarding. The work itself does not change.

Question 2

What do I need to run the Community Edition?

Accepted Answer

The source, and somewhere to run it. Rhesis is open source, so you can read it, extend it, and run it with Docker Compose on your own infrastructure. There is no key to request and nothing to sign, and when you run it yourself we set no limits on how much you use it. If you would rather not run it, a managed version exists. It is still in preview while we work out how it is offered, so talk to us and we will sort out your case together.

Question 3

What is the difference between self-hosting and Rhesis Cloud?

Accepted Answer

It is the same platform either way. With Docker Compose it runs on your own infrastructure. With Rhesis Cloud it runs on ours, at app.rhesis.ai. Nothing in the workflow changes between the two.

Question 4

Can I self-host the Enterprise Edition?

Accepted Answer

Yes. Rhesis is open source and self-hostable, and the Enterprise Edition adds access controls on top of it. It runs on your own infrastructure, or on ours.

Question 5

Do domain experts need a technical setup to take part?

Accepted Answer

No. Experts work in the web UI: they write cases in the Manual Test Writer, review results and traces, mark Pass or Fail with a comment, mention colleagues, and resolve threads. No code and no local setup.

Question 6

Do I need my own model API key?

Accepted Answer

No. The Rhesis Default model is hosted, so you can run your first evaluations without bringing a key. If you would rather use your own provider, LiteLLM routes to OpenAI, Anthropic, Gemini, Azure, Ollama, vLLM, and proxies.

Question 7

How do engineers get the annotated sets into CI?

Accepted Answer

Pull them with the Python SDK, the REST API, or MCP, then run them as a pipeline gate. There is a GitHub Actions guide, and the same pattern applies to GitLab CI, Jenkins, and CircleCI.

Question 8

Which agent frameworks are supported?

Accepted Answer

Connect any agent over REST or WebSocket, or register a Python function with the SDK Connector. Penelope simulates against Rhesis endpoints, LangChain, LangGraph, Microsoft Agent Framework, Pydantic AI, and custom agents, and tracing auto-instruments the same set.

Question 9

How do I move from Community to Enterprise?

Accepted Answer

Talk to us. There is no self-serve upgrade, because the access controls are set up together with your identity provider. Tell us how your organisation handles identity and we will take it from there.