Built for you.


Get in touch

We are currently in closed beta, working closely with our clients to finetune our API and ensure it is used in a responsible manner. You can join our waitlist: we will get in touch to learn more about your intended use and see if you are a fit for our closed beta.

Our API is currently available in French, Spanish, and Italian. To ensure high-quality, all our models are native speakers: they are trained from the ground-up in their target language. We are constantly adding new models: get in touch with us if you are interested in supporting a specific language.

Extreme-scale models are a natural fit for many challenging tasks across all kind of modalities: images, code, proteins, and more. We are actively working on expanding the Muse API beyond language. If you are interested, get in touch with us.

The models available in the Muse API are widely capable, and can be applied to most use cases directly through the endpoints we provide. If you feel like you are stretching the capabilities of our API, or if you have business-specific data you would like a model trained on, get in touch with us.

In the closed beta period, we work closely with each client to determine a pricing for their intended workload. After this period, we will have a simple and transparent pay-per-use offering.

Our API is built to deliver millions of words to thousands of end users. It is highly scalable, and can automatically meet your demand. If you think your use case could stretch its limits, please get in touch with us!

Whatever the level of NLP proficiency in your team, LightOn Muse will help your business.
If you don't use machine learning yet, our Process endpoints are a great way to get started immediately with common use cases. We can work with you to understand how our Muse API will best help you.
If you are already using NLP at scale, your pipeline will benefit from the rich representations available through our Understand endpoints. Find new use cases for your technology through the unique steerable text generation capabilities of create.