AI Usage & Privacy Filter

This page explains how Relyper uses AI (only for public or intended-public inputs), how the privacy filter works, and which obligations apply to external providers.

1. Processing Principles

• Public context: AI suggestions are generated from content that is publicly reachable or intended for publication.
• No default training on customer inputs: Providers must not use customer inputs for general model training by default (a separate opt-in would be required).
• Minimization & region: Processing uses minimal metadata and regional handling where available.

2. Privacy Filter (before AI)

Before sending to any AI provider, the snapshot goes through a privacy layer: pattern-based redaction (e.g., emails, phone numbers, addresses) plus contextual checks. The goal is to remove or minimize personal or sensitive data. This is best-effort; you remain responsible for legal compliance.

3. Provider Obligations

• No use of inputs for general training (without explicit opt-in).
• Limited or no retention of customer inputs.
• Regional processing where available.
• Secure transfer, access controls, auditability.

4. Transparency & Consent

By using the service, you consent to the processing described. We do not send non-public credentials or session cookies to external AI providers. See the Terms for details.

5. Results & Limitations

AI outputs are suggestions (e.g., tags, categories, summaries). They may contain errors. You decide whether to adopt them. For sensitive topics, we recommend manual review.