AI supports learning
NativePath AI features are designed for hints, explanations, code review, debug guidance and practice. They do not replace your own verification.
AI can be wrong, miss edge cases or suggest incomplete solutions. Review results before using them in real projects.
Data and limits
We try to send only the necessary task, code and lesson context to AI. Do not send secrets, private keys, tokens, customer personal data or production credentials.
AI usage may be limited by Pro access, credits, rate limits and safety checks.
How AI requests are processed
For AI hints, code review, debugging, and mentor chat, NativePath sends the request text and the necessary learning context to an external AI provider (currently OpenAI). That context may include user code, error messages, conversation history, and/or an excerpt from a published lesson. The reference solutionCode and exact validator-rule strings are not included in the AI context.
For usage accounting, diagnostics, and security, NativePath may store an account-linked, shortened and pre-redacted copy of the AI prompt and response. The current implementation limits each stored text to 4,000 characters and attempts to remove email addresses, phone numbers, API keys, and password/secret/token values. Do not send production credentials, private keys, tokens, or client personal data to AI features.