Privacy Policy

Poly was designed with privacy as a first principle. Your API response data never leaves your machine. We never see your payloads, user data, PII, business logic, auth tokens, or API keys.

The Poly SDK sends only schema metadata to Poly Cloud for drift analysis:

• Field names and data types (e.g., "name": "string")
• Schema structure (nesting, arrays)
• Drift type classification (rename, remove, type_change, etc.)

We never collect: actual field values, response payloads, user data, authentication credentials, or any PII.

Schema metadata is used exclusively for:

• Detecting API schema drift
• Generating field mappings via AI
• Improving detection accuracy

Schema metadata is stored in Supabase (PostgreSQL). We retain incident history for dashboard analytics. No response payloads or user data are ever stored.

We use Mistral AI for generating field mappings. Only field names and types are sent to Mistral — never values. We use Supabase for database storage.

Questions? hello@poly.dev