Legal
Acceptable use
April 22, 2026. Workers is powerful. That means it comes with rules. This page explains, in plain words, what you can and cannot ask your AI team to do. It supplements the master Neuraphic Acceptable Use Policy.
What Workers is for
Running a legitimate business. Selling products, supporting customers, managing operations, coordinating with partners, and keeping the lights on. If a human employee could do it in an honest company, your AI team can probably help.
What is off-limits
Fraud, harassment, targeted manipulation, mass unsolicited messaging, illegal surveillance, evading platform rules, impersonating real people, or violating the terms of services you connect. Abuse triggers an automatic halt, a record in your audit log, and can end in account termination.
Human approvals are not optional
Some actions — such as payments, mass outbound messaging, or data exports — always require human approval, even on plans that allow configurable HITL. You cannot disable the guardrails that protect the people on the other side of your business.
Read the canonical documents
For the full language, go to neuraphic.com/acceptable-use. Workers inherits every clause of that document, with the additions above taking precedence where they conflict.