Terms of use

Last updated: 14 November 2026

The service

WebGPU Chat is a free web application that downloads open language models and runs them in your browser. It is provided as is, without any guarantee of availability, accuracy or fitness for a particular purpose. It may change or stop being available at any time.

Responses can be wrong

The models are small and generate text probabilistically: they can make mistakes, invent facts, quotes or figures, or produce inappropriate content. You are responsible for checking the information and for how you use it. Do not rely on them as the only basis for medical, legal, financial or safety decisions.

Acceptable use

You must use the application lawfully and respect the licenses and usage policies of each model, which are linked on the licenses page. Do not use it to generate illegal content, to harass other people or to infringe third-party rights.

Your computer and your data

The models take up between 1 and 8 GB and make intensive use of the GPU while generating text, which can increase power consumption, temperature and fan noise. Your conversations are stored only in your browser: if you delete their data, they are lost and we cannot recover them.

Model licenses

When you download a model, you receive it from Hugging Face under its author's license: LFM Open License v1.0 for LFM3 (with limits on commercial use by entities with more than 10 million US dollars in annual revenue) and Apache 2.0 for Qwen3.5 and Gemma 4. See the licenses page.

Liability

To the extent permitted by law, the owner is not liable for damages arising from the use of the application or from the responses of the models.

Contact and applicable law

For any questions, write to amintahtahtoumi@gmail.com. These terms are governed by Spanish law.

This is a translation for convenience; the Spanish version prevails.