Skip to content
++++

Legal

Usage Policy

Last updated: September 19, 2026

This Usage Policy explains what you must not do with the Unbar inference API (the service). It applies to every account and every API key. An API key authorizes requests to the service. You accept this policy when you use the service. It is part of the agreement between you and us for the service.

Summary

You can use the service to build products and to run your own applications. You must not use it for the activities in this policy. These limits protect other people, protect the models that we serve, and keep the service available for everyone. If you break this policy, we can limit or end your access.

1. Illegal use

You must not use the service to break the law. This covers any activity that is illegal where you operate or where your users are. You must not use the service to plan, support, or carry out a crime. You must follow the laws that apply to you.

2. Child safety

You must not use the service to create, request, or share child sexual abuse material. You must not use the service to sexualize a child. You must not use the service in any way that endangers a child. When such material comes to our attention, we report it to the proper authorities.

3. Fraud and deception

You must not use the service to deceive or cheat people. This covers fraud, scams, and phishing. Phishing is a trick that makes a person reveal private information. You must not impersonate a person or an organization. You must not run deceptive or predatory schemes, such as false advertising or financial scams.

4. Malware and system attacks

You must not use the service to create or spread malware. Malware is software that harms a computer or its user. You must not use the service to gain unauthorized access to a system, an account, or a network. You must not use the service to attack or disrupt computer systems, networks, or critical infrastructure. Critical infrastructure includes power, water, health, and communication systems.

5. Weapons

You must not use the service to develop or design weapons. This covers biological, chemical, nuclear, and radiological weapons. It also covers conventional weapons and their key parts. You must not use the service to help produce or deploy a weapon.

6. Violence and hate

You must not use the service to plan or promote violence. You must not use the service to support terrorism. You must not use the service to incite violence against a person or a group. You must not use the service to incite hatred against a person or a group.

7. Privacy of other people

You must not use the service to violate the privacy of other people. You must not use the service for mass surveillance. You must not collect personal data without a lawful basis. A lawful basis is a legal reason to process personal data. You must not track a person without a lawful basis. Personal data is information that relates to an identified or identifiable person.

8. Spam and disinformation

You must not use the service to send spam. Spam is bulk unwanted messages. You must not use the service to run coordinated disinformation campaigns. Disinformation is false information that a group spreads on purpose. You must not use the service to manipulate an election or a democratic process.

9. Abuse of the service

You must not abuse the service itself. You must not share or resell your API keys against your agreement with us. You must not evade rate limits or usage limits. A rate limit caps how many requests you can send in a period. You must not probe, scan, or attack the platform. You must not try to break the security of the service.

10. Your responsibility

You are responsible for how you use the service. You are responsible for the prompts that you send. A prompt is the input that you send to a model. You are responsible for the completions that you generate. A completion is the output that a model returns. You are responsible for how you use these completions.

11. Enforcement

We enforce this policy when a violation is reported to us or otherwise comes to our attention. We do not monitor, scan, or inspect your prompts or completions. We do not store this content, and we do not read it. So we cannot detect a violation inside your prompts or completions on our own. We also act on information from operating the service, such as billing and usage records. These records do not contain your prompts or completions.

We reserve the right to throttle, suspend, or end access for a violation of this policy. To throttle is to slow the rate of requests. We choose a response that fits the violation. We do not block or change the completions that a model returns to you.

12. Report abuse

If you see abuse of the service, tell us. Report abuse to support@unbar.sh. Describe the problem and give any account or request details that you have. We review reports and act where this policy applies.

13. Changes

We publish changes to this policy on this page with a revised date.

++