TestZeus Docs

The testing layer for coding agents

Plug TestZeus into your agent as a testing system.

Author, run, and report on tests in plain language — straight from the CLI, the Python SDK, the MCP server, or as agent skills. Describe what to check; TestZeus does it.

Get started → Which surface should I use?

How it works

01

Author

Describe a test in plain language or Gherkin — no scripting.

02

Run

TestZeus executes the test and captures every step.

03

Report

Get pass/fail results, logs, and artifacts back in your workflow.

Four surfaces, one platform

Pick how you wire it in

Compare all four →

CLIpip install testzeus-cli Terminal, shell scripts, and CI pipelines — scriptable, with exit codes CI can gate on. CLI reference → Python SDKpip install testzeus-sdk A typed async client for building TestZeus into your own apps, services, and automation. SDK reference → MCP servernative tools Claude Desktop and Cursor call TestZeus as native tools — run and inspect tests in conversation. Add the MCP server → Skillsfor coding agents Plain-English playbooks Claude Code / Codex load on demand to author and run whole suites. Browse skills →

Always in sync — these docs are auto-generated

The CLI, SDK, MCP and Skills references are generated by parsing the source repos, so every command, method, tool and skill here matches the code. The site rebuilds nightly and on every source change — only the overview and getting-started pages are hand-written.