AI Model Training Policy

Last updated: 2026-06-24. Your data is never used to train AI models — not ours, and not our AI sub-processors'. We only send the minimum content needed to operate the service, and only for live inference.

Our commitment

Content you provide to qa-clone — page snapshots, test prompts, run artifacts, and the natural-language goals you write — is used solely to operate the service: generating tests, executing runs, classifying failures, healing flaky steps, and reporting results. It is never used to train, fine-tune, or improve any machine-learning model.

AI sub-processors

We call third-party large-language-model and embedding APIs to power these features: Anthropic for test generation, failure classification, and healing, and OpenAI for the embeddings behind page memory. Per their published API terms, neither Anthropic nor OpenAI uses data submitted through their APIs to train their models.

What we send

We send only the minimum content a feature needs at the moment it runs — for example, the relevant page text and your test prompt — and nothing more. We do not export your data to AI providers in bulk.

Verification

The providers we use and the data categories each one receives are published on our Sub-processors page, and every sub-processor operates under a data processing agreement. The major providers' API data-usage policies are public: API inputs and outputs are not used to train their models.

See the Sub-processors page for the data each provider receives, and our Privacy Policy for how we handle personal data overall.