Acceptable Use Policy

Last updated: July 7, 2026

AlchemyLake exists to turn governed data into trustworthy deliverables. These rules keep it safe. They apply to every surface — Studio, REST API, MCP endpoint, and the Databricks app. Prompts and outputs are screened by a moderation layer; blocked runs are refunded.

Never allowed

• Sexual content involving minors — always blocked, always reported where required by law.
• Non-consensual sexual content or sexualized real persons.
• Instructions or material assistance for weapons, explosives, or biological, chemical, or nuclear harm.
• Credible incitement of violence against real people or groups.
• Malware, credential theft, phishing kits, or network-intrusion tooling.
• Deceptive synthetic media presented as authentic — including stripping AlchemyLake’s provenance marks to pass AI output off as human or documentary content.
• Uploading data you have no right to process (stolen datasets, scraped personal data without a lawful basis).

Not allowed without care

• Regulated advice (medical, legal, financial): deliverables may inform, not replace, licensed judgment — you must add human review before use.
• Political persuasion at scale: clearly label AI-generated material and comply with local election law.
• Third-party brands and likenesses: you are responsible for having rights to logos, faces, and trademarks you include in prompts, uploads, or brand kits.

Platform integrity

• No attempts to bypass safety moderation, rate limits, or credit metering.
• No probing, scanning, or disrupting the Service or other tenants.
• Automated use is welcome through the documented REST/MCP surfaces with your own API key — not by scripting the web UI or sharing session tokens.

Enforcement

Violations may lead to withheld outputs (refunded), throttling, suspension, or termination depending on severity, and disclosure where the law requires it. Appeals: info@zorost.com.