Acceptable Use Policy

Last updated: 29 August 2026

This policy forms part of the Terms of Service. It exists because the Service reads the public web on your behalf, and that capability has to be used responsibly.

1. What the Service is for

Programmatic web search, retrieval and content extraction, for use in applications, AI agents, research and retrieval augmented generation.

2. You must not use the Service to

• Break any applicable law, or to help anyone else do so.
• Infringe copyright, database rights, trade marks or any other third party right, including by systematically reproducing or republishing substantial parts of a source's content as your own product.
• Collect personal data unlawfully, build profiles of private individuals, or conduct surveillance or stalking of any person.
• Circumvent authentication, paywalls, access controls or technical protection measures on any site. The Service retrieves publicly accessible pages only.
• Attack, overload, disrupt or denial of service any third party site or the Service itself, or use it as a component in such an attack.
• Harvest contact details for unsolicited bulk messaging.
• Generate or distribute malware, phishing content, or material that facilitates fraud.
• Produce sexual content involving minors, content that incites violence or hatred, or content promoting self harm.
• Create deepfakes or impersonate a real person or organisation deceptively.
• Resell raw access to the Service as an undifferentiated search API without written agreement from us. Building a product on top of it is exactly what it is for.

3. Respect the sources

The Service returns third party content. That content belongs to whoever published it. When you use it, attribute and link the source where you display it, do not present someone else's writing as your own, and honour any licence attached to it. Every result the Service returns includes its source URL precisely so this is possible.

4. Fair use of capacity

Rate limits are published in the API reference. Do not attempt to evade them by rotating keys, and do not create multiple keys to extend the free allowance. Sustained load beyond your limits may be throttled. If you need more capacity, ask us rather than working around it.

5. Content generated by the Service

Answers are generated by a language model from retrieved passages. They can be wrong. Where the Service marks a supporting quote as unverified, treat the claim as unverified. Do not present generated answers as professional legal, medical or financial advice.

6. Enforcement

We may investigate suspected breaches and may suspend or terminate a key. For a clear or serious breach we may act without notice. Where we suspend a key for reasons other than your breach, we refund unused credits.

7. Reporting

To report abuse of the Service, or content returned by it that you believe infringes your rights, email [email protected] with the details and we will respond within three business days.