Acceptable use policy

1. Scope

Applies to everyone using the dashboard, API, plugins, loader, public endpoints and published bundles under your account.

2. You must not

1. Publish unlawful, infringing, deceptive or harmful content through our storage/CDN. Bundles are public (Terms 6.2). OPEN: categories (hate, adult, malware, phishing, counterfeit, etc.) for counsel.
2. Put personal data, secrets, credentials or confidential text into text you translate and publish. Source text goes to third-party engines and the result is published on public CDNs.
3. Scan or crawl sites you do not own or control. The crawler is SSRF-guarded and honours robots.txt unless you prove ownership. Do not try to use it to reach internal networks, cloud metadata or other private addresses.
4. Abuse the public discover endpoint (site-key plus allowed-origin string reports): no flooding, no spoofed origins, no injecting hostile text. Rate limits: per key 600/min, per IP 1,200/min, discover 60/min per IP.
5. Circumvent limits: sites, languages, word allowances, rate limits, trial limits, or licence activation limits; create accounts to evade a suspension; use disposable or false e-mail addresses to sign up.
6. Share or resell access: sharing secret API keys or licence keys outside your organisation; reselling the service as a competing translation service. OPEN: Agency white-label terms.
7. Attack the service or other customers: attempt to bypass authentication, tenant isolation, signature verification or the preview-origin rules; probe for other tenants' data; deliver malware; overload the platform.
8. Misuse machine translation: attempt to extract engine providers' models or credentials, prompt-inject the LLM engine to produce unrelated output or exfiltrate prompts, or use the service to bulk-translate third-party content you have no right to use.
9. Send unsolicited e-mail through webhooks, invitations or any feature, or register webhook endpoints that point at systems you do not control (the platform blocks private and local addresses:).
10. Impersonate a person or organisation, or falsely imply our endorsement.

3. Security research

Good-faith testing of your own account is welcome within the limits above. Testing of production systems belonging to us or to other customers needs written permission. OPEN: a vulnerability-disclosure policy and contact (none exists), safe-harbour wording verify.

4. Enforcement (what the platform can actually do)

• Operators can suspend or unsuspend a tenant or a site, cancel jobs, decide on quarantined discover submissions in the abuse review queue, and switch off discover per site; every action needs a reason and is audited; tenant and site suspension appear in the customer's audit log.
• There is no tool to take down already-published bundles from CDNs on abuse or legal grounds. Until one exists, takedown is a manual storage operation. This is a gap counsel must weigh.
• Notice, appeal and warning steps before suspension: OPEN. Immediate suspension for security or legal risk: OPEN.

5. Reporting and legal notices

Abuse contact and copyright-notice procedure: OPEN (notice-and-takedown regimes such as the US DMCA safe harbor or the EU Digital Services Act apply differently to hosts, caches and intermediaries: verify, OPEN whether we count as a hosting service).