Usage & Costs
Testronaut itself is free to run β but it uses your own LLM API key behind the scenes.
This means that any tokens consumed during missions are billed directly to your LLM provider account (OpenAI, Gemini, etc.), not to Testronaut.
- You bring your own LLM API key
- Missions consume tokens (reasoning + DOM handling)
- Your provider bills you directly for those tokens
- Testronaut adds no markup or extra fees
How costs workβ
- Generate an API key from your chosen LLM provider.
- Add your key to
.envor your project config. - Run missions β each turn consumes tokens for reasoning, tool calls, and DOM processing.
- Your provider bills you directly based on usage and published pricing.
Monitoring usageβ
OpenAIβ
- View usage and spend in your OpenAI dashboard.
- Set spending limits under Billing β Usage limits.
- CLI runs show tokens used per step for transparency.
Geminiβ
- Track usage in your Google AI Studio or Cloud Console.
- Set quotas or rate limits under your Cloud project.
- Gemini responses also include usage metadata (
totalTokenCount).
Example CLI outputβ
[model] β navigate login page
[tool ] β β
Success
Tokens used this turn: 628 (β $0.0025 at GPT-4o-mini or Gemini-Flash rates)
Estimating costsβ
Actual cost depends on:
- Provider and model (e.g. GPT-4o vs Gemini-2.5-Pro)
- Mission length, retries, and DOM size
- Whether screenshots or reasoning steps are requested
Typical ballpark:
| Scenario | Approx. Cost |
|---|---|
| Simple login flow | A few cents |
| Multi-page checkout | Tens of cents |
| Long exploratory mission | Up to a few dollars |
βCreditsβ in documentationβ
References to credits mean token usage billed by your provider β
there is no internal billing system inside the open-source Testronaut CLI.
Controlling spendβ
- Use lighter models (
gpt-4o-mini,gemini-2.5-flash) for lower-cost runs. - Keep missions focused and limit retries.
- Set provider-level usage caps or quotas.
- Monitor dashboards regularly for visibility.
FAQβ
Do I pay Testronaut directly?
No. Testronaut doesnβt charge you; you pay your LLM provider.
Do I need my own LLM account?
Yes β you must supply your own API key (e.g., OpenAI or Gemini).
Can I limit usage?
Yes β both OpenAI and Gemini let you set limits and budgets in their dashboards.
Does Testronaut add any markup?
No. All costs are at standard provider rates.