Acceptable use
Your requests reach model providers through minirouter’s own upstream accounts, shared by every user. Abuse that gets one banned degrades the service for everyone.
Prohibited
- Child sexual abuse material
- Absolute, in any form, generated or uploaded. Reported where the law requires.
- Attacks on people
- Harassment campaigns, doxxing, deanonymization of private individuals, credible threats.
- Fraud at scale
- Phishing kits, romance-scam automation, fake-review farms, impersonation of real people or organisations.
- Malware and intrusion
- Generating or operating tooling to compromise systems you do not own or have authority to test.
- Circumventing upstream policy
- Using our routing to evade a ban, quota or safety policy a provider has applied to you.
- Anything illegal where you are
- The catch-all is yours to interpret conservatively; we are not your counsel.
Each upstream provider additionally applies its own usage policy to traffic it serves. Where a provider’s policy is stricter than this list, their policy governs requests routed to them — we cannot and will not launder traffic past it.
How this is enforced
Request content is retained for debugging as described in our privacy policy. Enforcement also uses the following signals: abuse reports from upstream providers about our account traffic, anomalous usage patterns, payment-screening hits, and reports sent to us directly. Segregating suspicious traffic onto isolated upstream accounts is our first response, precisely so that a provider’s countermeasure lands on the abuser’s traffic instead of on everyone’s.
Keys in breach are suspended, and the suspension is explicit — the API returns a named error with a contact route rather than a silent failure. Because refunds are service credit only, suspension of an abusive key is a real cost; do not prepay a balance you intend to burn on the list above.
To report abuse of this service — including abuse aimed at you — write to abuse@minirouter.sh. A human reads it.