FAQ ❔

Does Testronaut need a headless browser?
Yes, it runs in headless mode by default. You don’t need to manually open a browser window unless you enable debugging or visualization.

How do I track the token consumption of my tests?
Each mission is limited to 20 turns by default. Token usage per turn is shown in your mission reports, helping you understand how efficiently the agent completed each step.

How do I take a screenshot?
Add a natural instruction like:

Take a screenshot.

to your mission prompt text whenever you want the agent to capture the current state. Screenshots automatically appear in the test reports and are saved under:

missions/mission_reports/screenshots/

What LLMs are supported? Testronaut supports multiple LLM providers, letting you choose the one that best fits your workflow.

Currently supported:

• OpenAI — GPT-4.1, GPT-4o, o3, o4-mini, and GPT-5 families

• Google Gemini — Gemini 2.5 Pro, 2.5 Flash, and 2.5 Flash-8B

You can select your preferred provider during initialization (npx testronaut --init), or switch later by editing your testronaut-config.json.

⚠️ Note: Availability and rate limits depend on your API account and region. Some higher-tier models (e.g. GPT-5) may be restricted or rate-limited by the provider.