Getting started#
This guide takes you from zero to a running TestZeus test. TestZeus is the testing layer your coding agent — or your own code — can drive: author tests in plain language, run them, and get results back.
Prerequisites
A TestZeus account (sign up at https://prod.testzeus.app) and Python 3.11+.
1. Install & authenticate#
pip install testzeus-cli
testzeus login
You’ll be prompted for your email and password; credentials are stored securely in your system keyring. Keep separate logins with testzeus login --profile dev.
2. Run your first test#
Pick the surface that matches how you work — they all hit the same API (https://prod.testzeus.app/api).
testzeus tests list # see your tests
testzeus tests get <id> # inspect one
Best for quick checks and CI. See the CLI reference.
import asyncio
from testzeus_sdk import TestZeusClient
async def main():
async with TestZeusClient(email="you@example.com", password="...") as client:
tests = await client.tests.get_list(sort="id")
print(tests["items"][0].data)
asyncio.run(main())
Best for building TestZeus into your own app. See the SDK reference.
Point Claude Code / Codex / Cursor at the TestZeus skills, then ask in plain English:
“Create a login test for my app and run it in staging, then show me the results.”
The agent authors and runs the test for you. See Skills.
Add the TestZeus MCP server to Claude Desktop or Cursor and the tools appear natively. See MCP Tools.
3. Pick the right integration#
Not sure which surface to use? The Use cases page breaks down when and why to choose the CLI, the SDK, the MCP server, or skills — with an example for each.