Terms of Use

1. What OpenShore is

OpenShore is a tool for building software with models you choose. You can run models on your own hardware, or connect a cloud provider on your own account and your own key. Your work stays yours.

The OpenShore source code is licensed under the Apache License 2.0. These terms cover your use of the apps we ship and our account service, and do not limit what that license allows.

2. Prohibited uses

You may not use OpenShore to generate, describe, plan, refine, or assist with:

• Child sexual abuse material

  Sexual content involving minors, in any form, including drawn, written, and synthetic depictions.

• Non-consensual intimate imagery

  Undressing or "nudifying" an image of a real, identifiable person, placing a real person's face into sexual material, and any other sexualization of an identifiable person who has not consented.

• Weapons capable of mass casualties

  Actionable assistance with synthesizing or deploying biological, chemical, nuclear, or high-yield explosive weapons, or with materially increasing the harm such a weapon would cause.

• Cloning a real person without authorization

  Synthesizing the likeness or voice of a real, identifiable person, as an image, a video, or an audio clip, is permitted only when you are authorized for that specific person: they are you, they gave you permission, or they licensed the likeness to you. You assert that authorization in the product, and the assertion is recorded on your device; it is never sent to us. This is about synthesized media, not words: writing about, satirizing, or criticizing a person in text is not covered here.

The first three have no exception and no consent option. The fourth is gated behind your authorization, not forbidden.

3. What is not prohibited

To be clear about the boundary, because a vague rule chills legitimate work. Legal adult content between adults, dark and violent fiction, horror, and edgy humor are permitted. Satire, parody, and criticism of public figures are permitted. Security research, red teaming, exploit analysis, and vulnerability work for defensive purposes are permitted. Controversial, unpopular, and dissenting opinions are permitted.

Words are free here. The consent gate in section 2 is on synthesized media, not on what you write: satire and parody in text never touch it. "It's satire" does not, however, clear a photorealistic image or video of a real person's face, because a synthesized face is taken as real until proven otherwise, and that asymmetry is the whole problem.

OpenShore does not add refusals or commentary to any of this.

4. Consequences of misuse

Misuse under section 2 leads to permanent removal from the platform. A Tier 1 attempt flags the account for termination and prepares a report; the removal is carried out by the operator, and where the law requires or permits it the matter is reported to the appropriate authority. In the United States, child sexual abuse material is reported to the National Center for Missing and Exploited Children. Account termination and, where required, a report are the entire enforcement scope. We never store an IP address in our database or use one for enforcement. Our hosting providers see it in transit and keep short operational logs, as any web service does; the privacy policy names them.

We take this seriously and we want to be direct about why. Deepfakes, non-consensual intimate imagery, and synthetic child sexual abuse material are spreading because the tools to make them became easy to get. This product exists in part to be one of the tools that does not help, and enforcement is how that promise is kept rather than merely stated.

5. What we record when something is blocked

We record the category, the tier, the time, which model path served it, and a keyed one-way hash (HMAC-SHA-256) of the request. We keep it for 180 days, longer only under a legal hold.

We do not store the prompt, the answer, or any excerpt of either. A hash is enough to recognize a repeat and to identify the same content in a lawful report, and retaining harmful material in order to police harmful material is its own harm.

6. Honest limits

OpenShore enforces these boundaries as shipped, and does not help you remove them. There is no setting, configuration file, or environment variable in the product that turns the guardrail off.

We cannot control models you run independently on your own hardware. Open model weights on your own machine are outside the reach of any application, including this one. We do not claim otherwise, and we do not claim that misuse is impossible. What we guarantee is narrower and real: this app, as shipped, does not assist the uses in section 2 and does not help you strip these protections out. Because the code is open, anyone can change their own copy; these guarantees cover the builds we ship, not modified ones.

7. Standards alignment

OpenShore aligns its risk practices with the NIST AI Risk Management Framework and ISO/IEC 42001, and attaches content provenance using the C2PA assertion vocabulary.

This is a self-attestation of alignment. No third party has certified, endorsed, or audited this product against those frameworks, and we will not say otherwise. Provenance records written by this product are unsigned, because signing requires a certificate from a C2PA-recognized authority that OpenShore does not currently hold.

8. Your content and your models

Your prompts, code, and generated work are yours. Local model work stays on your machine, except the words of a web search a model makes while web search is on, a web page the coding agent opens with your say so, and what a cloud model in the same chat reads on a turn it answers. Cloud model work goes to the provider you connected, on your account, under that provider's terms. A cloud provider's own policies apply in addition to these terms, never instead of them.

9. Changes and contact

We will update these terms as the product changes. Material changes to section 2 or section 4 will be surfaced in the app.