Unfenced - Acceptable Use Policy
Last updated: 17 September 2026
This Acceptable Use Policy ("AUP") is part of the Terms of Service and applies to everyone who uses Unfenced. It exists because Unfenced can do something most tools cannot: act on a third-party website while signed in as you. That power comes with rules.
If you break this policy, we may suspend or terminate your access (Section 6).
1. The rule that matters most: authority
Unfenced acts on other people's services under your authority. Before the Service will act on a site, you must add it to your owner-controlled act-allowlist - the Service will not act on a site you have not authorized, and an agent cannot silently grant itself a new site.
You are responsible for everything your agent does through the Service, exactly as if you had done it yourself (see Terms of Service, Sections 3 and 5). You may only authorize, and only direct the Service to act on, sites and accounts where you have the right to do so. Specifically, you must not use Unfenced to:
- access an account you do not own and are not authorized to operate;
- use someone else's credentials without their authorization;
- circumvent a login, paywall, CAPTCHA, rate limit, or other access control that is not yours to circumvent; or
- continue accessing a site after its owner has told you (or us) to stop.
2. No unlawful or terms-violating access
You must not use Unfenced to:
- break the law - including computer-misuse / anti-hacking laws, data-protection laws, intellectual-property laws, sanctions, or export controls;
- violate a site's terms - scraping or automating a site in breach of its terms of service, its
robotsdirectives, or its API terms, except where those terms are unenforceable as a matter of law; - evade blocks in bad faith - rotating identities, or using our real-browser and real-IP capability, specifically to defeat a site's decision to refuse automated access.
Unfenced returns a structured blocked result when a site refuses us. Treat that as the site's decision. Do not use the Service to organize a campaign to overcome it against the site's wishes.
3. No misuse of credentials or personal data
- Do not store, in the credential vault or a saved session, any credential you do not have the right to use.
- Do not use the Service to harvest personal data you have no lawful basis to collect, or to build profiles of people without a lawful basis.
- Do not attempt to extract stored secrets. There is no reveal route by design; do not probe for one, and do not try to trick an agent into echoing a secret.
- Respect the privacy of third parties whose data appears in pages you fetch or sessions you drive. When you process their data through the Service, you are the controller and are responsible for having a lawful basis.
4. No targeting of individuals
Do not use Unfenced to:
- harass, stalk, dox, surveil, or intimidate a person;
- assemble a dossier on a specific private individual without a lawful basis;
- scrape or monitor a person's accounts, posts, or profiles against their wishes; or
- facilitate discrimination or any use that endangers a person's safety.
5. No abuse of the Service or the wider web
You must not:
- send volumes of traffic through the Service designed to overload, degrade, or deny service to a third-party site (no DoS/DDoS, no abusive crawling);
- exceed rate limits we set, or engineer around metering, throttling, or abuse controls;
- use the Service to send spam, phishing, or malware, or to distribute unlawful, infringing, or harmful content;
- probe, scan, or attempt to breach the security of the Service or its infrastructure, or of any system you are not authorized to test; or
- resell or provide the Service to third parties except as your Terms permit.
Rate and abuse limits. We meter usage (fetches, session time, and similar) and may apply rate limits, quotas, and automated abuse detection. The limits in force during the preview are:
| Limit | Value |
|---|---|
| Requests per account | 600 per minute |
| Requests without an account | 1,200 per minute per address |
| Job status polling | 6,000 per minute |
| Beta waiting-list submissions | 300 per minute per address |
| Fetches per account per day | 2,000 |
| URLs in one batch | 20 |
| Words returned per batch | 25,000 |
| Idle live page before it closes | a few minutes |
| A parked live page | up to 30 minutes |
Exceeding a limit returns a refusal (quota-exceeded or server-busy), not a silent truncation. These are preview values and we may change them at any time without notice - they exist to protect the Service and the sites it reaches, not as an entitlement. Sustained attempts to evade them are a breach of this policy.
6. Enforcement
- We may investigate suspected violations and cooperate with law enforcement.
- We may remove content, disable a credential or session, throttle, suspend, or terminate access - with notice where practical, and immediately where a violation is serious, ongoing, or legally required to stop.
- We may act on well-founded abuse reports from third parties (see
abuse-takedown.md). - Violations that expose us or others to legal risk may also trigger the indemnity in the Terms of Service.
7. Reporting abuse
If you believe someone is using Unfenced in violation of this policy, report it via abuse-takedown.md or at abuse@unfenced.ai.
_This AUP supplements, and is incorporated into, the Terms of Service._