Terms of Use

OpenGradient Chat is an interface that lets you send prompts to third-party AI models (such as ChatGPT, Claude, Gemini, and Grok) through a privacy-preserving gateway. This includes text conversations, image generation, the Video Studio — a feature that generates, animates, upscales, and edits video and audio from your prompts — and the Local Agent — a feature that writes and runs code in a sandbox on your own device. The underlying output is generated by the upstream model you select. We do not control or warrant the accuracy, reliability, or safety of model responses, generated images, generated video or audio, or code produced or executed by the Local Agent.

All metadata that contains personal information (for example, your IP address) is removed before sending Prompts to the underlying model providers (for example, Anthropic, OpenAI, Google, and xAI). If you submit personal information in your Prompts, it may be reproduced in the Outputs, but no one can tell (including us and the underlying model providers) whether it was you personally submitting the Prompts or someone else.

Your chat history is encrypted in your browser with a key only your device holds. We do not store your conversations on our servers, and we cannot read them. If you clear your browser storage or sign in from a new device without your key, the history is unrecoverable.

You should also know that some browsers automatically delete locally-saved chats if you have not accessed OpenGradient Chat within so many days, while others retain them for extended periods or indefinitely, unless you explicitly clear your browser data. You can opt out of saving recent chats altogether at any time.

You must be at least 13 years old (or the minimum age in your jurisdiction) to use the service. You are responsible for activity that occurs under your account and for keeping your credentials and local encryption keys safe. If your device or browser storage is lost, your local chat history may be unrecoverable — we cannot reconstruct it on your behalf.

OpenGradient Chat allows you to use models from leading model providers without being tracked. These model providers have usage policies, and our Terms therefore mirror those policies. For example, usage policies from Anthropic, OpenAI, Google, and xAI apply to their respective models.

Accordingly, it is a violation of these Terms to use OpenGradient Chat for anything:

• Illegal, discriminatory, or harmful, including generating hateful, harassing, or threatening content; terrorism, violent extremism, or animal abuse content; or content promoting suicide, self-harm, or disordered eating; or using Outputs for anything fraudulent or deceptive.
• Violating others' rights or property, including privacy and intellectual property rights, creating spam, malware, computer viruses, or gaining access to computer systems without authorization.
• Sexually explicit or obscene.
• Endangering children, including generating or uploading child sexual abuse material (CSAM), including AI-generated CSAM; facilitating the exploitation or abuse of minors; or sexualizing minors in any context. [Note: We report CSAM to the relevant authorities.]
• With high risk of physical or economic harm, including designing weapons, explosives, or other dangerous materials, or promoting multi-level marketing, predatory financial practices, gambling, or sports betting.
• Generating photorealistic images, videos, or audio of real people without their consent, including deepfakes, non-consensual intimate imagery (NCII), or content that could mislead others about authenticity.
• In a regulated area, including providing legal, financial, or medical advice or services, political campaigning or lobbying, determining eligibility for financial products or creditworthiness, educational admissions or testing, housing (leases and home loans), employment, or more generally for governmental decision-making, such as law enforcement or immigration decisions, or for facilitating academic dishonesty.
• Interfering with or negatively impacting OpenGradient Chat or other OpenGradient services, including circumventing rate limits, safeguards, or other restrictions, or automated querying and developing or offering AI services, such as training an AI model.

We may suspend or terminate accounts that violate these terms or that put the service or other users at risk.

To avoid misuse, safety mechanisms — both ours and those of the upstream model providers — may block certain Prompts and the generation of certain Outputs.

OpenGradient Chat can generate images from your Prompts using third-party image models. Generated images are Outputs, and the acceptable-use restrictions above apply to them in full — in particular, you may not generate child sexual abuse material, non-consensual intimate imagery, or photorealistic images, video, or audio of real people without their consent.

AI-generated images may unintentionally resemble real people, existing works, trademarks, logos, or other protected material. You are solely responsible for reviewing the images you generate and for ensuring that your creation and use of them does not infringe anyone's intellectual property, publicity, privacy, or other rights, and does not mislead others about authenticity. We make no warranty that generated images are original, non-infringing, or suitable for any purpose, and you use them at your own risk.

The Video Studio lets you generate, animate, upscale, and edit video — some of it with synchronized audio — from your Prompts, from images you upload or generate, and from templates, using third-party video and audio models. Generated clips are Outputs, and every acceptable-use restriction above applies to them in full. In particular, you may not create child sexual abuse material, non-consensual intimate imagery, or photorealistic video or audio of real people — including their voice or likeness — without their consent, and you may not use the Video Studio to produce deepfakes or other content that could mislead others about authenticity.

Some video models available in the Video Studio are open-weights models hosted without provider-side prompt moderation. The absence of an automated filter does not permit content that these Terms or applicable law prohibit: the acceptable-use restrictions apply to those models exactly as they do to every other model, and you remain fully responsible for the Prompts you submit and the clips you generate. We may suspend or terminate accounts that use these models to create prohibited content.

AI-generated video and audio may unintentionally resemble real people, voices, existing works, trademarks, logos, or other protected material. You are solely responsible for reviewing what you generate and for ensuring that your creation, download, and use of it does not infringe anyone's intellectual property, publicity, privacy, or other rights, and does not mislead others about authenticity. We make no warranty that generated video or audio is original, non-infringing, or suitable for any purpose, and you use it at your own risk. Generated clips are stored locally in your browser and may be lost if that storage is cleared.

The Local Agent is an experimental feature that autonomously writes and runs code, reads and writes files, researches the web, and builds documents and apps inside a sandbox that runs in your browser, on your own device. Your files and the code it runs stay on your device; only the model requests that drive the agent leave, and those are anonymized through our gateway as described above.

The agent runs only inside this in-browser sandbox: it works with the files in its workspace and cannot reach or alter other files on your device, and its code-execution environment has no network access. Even so, the agent acts autonomously, and the code and results it produces may be inaccurate, insecure, or behave in unintended ways. You are responsible for supervising it and for reviewing its code and output before relying on them or running them outside the sandbox. Files and results held in the workspace live in your local browser storage and may be lost if that storage is cleared.

To the maximum extent permitted by law, we are not responsible for inaccurate results, for any code the Local Agent generates, or for any consequences of your using or running that code or those results outside the sandbox. The sandbox limits what the agent can reach, but it is not a guarantee of correctness or security.

If you believe that your intellectual property rights have been infringed, please email legal@opengradient.ai with the details of your claim.

To report Outputs that you believe violate these Terms of Use or the law, you can report them to us using the feedback option in the chat interface (thumbs-down, with the “Harmful or illegal” option selected) or by emailing legal@opengradient.ai.

Users in the EEA may contact legal@opengradient.ai to access our internal complaints handling process, as defined in the EU Digital Services Act.

You agree to indemnify and hold harmless Vanna Laboratories, Inc., its affiliates, employees, and any other agents from and against any claims, losses, and expenses (including attorneys' fees) arising from or relating to your use of OpenGradient Chat, including your subsequent use of any Outputs, your breach of these Terms, or violation of third-party rights or applicable law.

You retain all rights to the prompts you submit (“Input”) and the responses returned to you (“Output”), subject to the rights of the underlying model providers. You are responsible for ensuring you have the rights to submit your Input and for how you use any Output.

Because chat history is encrypted on your device, we have no copy of your Input or Output. You grant us only the limited rights necessary to route your request to the provider you selected.

Some features require credits or a paid plan. Prices and any applicable taxes are shown at checkout. Credits are consumed as you use the service and may expire as disclosed at purchase. Except where required by law, payments are non-refundable. We may change prices for future purchases on reasonable notice.

When you select a model, your prompt is forwarded to that provider through our gateway. Your use of those models is also governed by the provider's own terms and policies. We are not responsible for provider availability, content, or changes to their offerings.

To the maximum extent permitted by law, the service is provided “as is” and “as available” without warranties of any kind, whether express or implied, including warranties of merchantability, fitness for a particular purpose, non-infringement, or accuracy of model output, generated images, generated video or audio, or agent-generated code. AI models can produce inaccurate, biased, or misleading information; generated images, video, and audio may resemble real people, voices, or protected works; and code written or executed by the Local Agent may contain errors or behave unexpectedly. You should not rely on Output for medical, legal, financial, or other professional advice.

To the maximum extent permitted by law, Vanna Laboratories, Inc. and its affiliates will not be liable for any indirect, incidental, special, consequential, or punitive damages, or any loss of profits, revenue, data, or goodwill, or any damage to your device, files, or systems (including from your use of code generated by the Local Agent), arising out of or related to your use of the service. Our aggregate liability for any claims arising out of or related to the service will not exceed the greater of (a) amounts you paid us in the twelve months preceding the event giving rise to the claim and (b) one hundred U.S. dollars.

You may stop using the service at any time. We may suspend or terminate your access if you violate these terms, if we are required to do so by law, or if we discontinue the service. On termination, the sections of these terms that by their nature should survive will survive.

We may update these terms from time to time. When we make material changes, we'll update the “Last updated” date above and, where appropriate, notify you in the app or by email. Continued use of the service after changes take effect constitutes acceptance of the updated terms.

Questions about these terms? Email legal@opengradient.ai.