Not Diamond

The perfect answer every time.

‍One simple API to drive down cost and latency while boosting quality and reliability.

Maximize quality

We route to dozens of models to maximize response quality.

Data-driven learning

Easily provide feedback data to optimize model routing to your use case.

No more eyeballing

Leverage precise, data-driven decisions that continuously improve.

Route to anything

Route to custom models, workflows, agents, chains, or anything else.

Try it out yourself.

Send some queries below and watch them get dynamically routed to the best-suited model.

Always high-quality.

By default, we always route to the most accurate model available, ensuring your application is consistently high-quality.

You're in control.

Easily define personalized routing logic, integrate with your custom workflows, and manage client-side API calls to models.

The perfect fit.

Quality is subjective. Easily provide feedback data on routing decisions and we’ll instantly adapt to your users' needs.

An AI safety and research lab.

We've built venture scale companies, pushed code to millions of users, and published cutting edge research in top AI journals. We believe that the interfaces between models are just as important as the quality of the models themselves—not only for performance and generalization, but for safety as well.

‍Read more about our vision for long-term safety.