Focus on your business logic and application while letting Glide handle common LLMOps problems (like fallbacks, cost control, routing, caching, etc.)

High performance and scalability are our priority, so we don’t slow you down and scale out as you grow.

Glide is designed around resiliency and failures, so your application can achieve high availability working with external models and survive the toughest real-world situations.

A Glide cluster can serve multiple applications with totally different needs.

Switch your model providers seamlessly for your application without a need to modify them.

We welcome all kinds of feedback and contributions in Glide. Let’s make it work in the best possible way for you!

Change your model providers without a need to modify your applications

Written in Golang with the best practices of software development in a cloud-native environments in mind

Flexible routing across several models and various strategies like the least latency, weighted round-robin, round-robin, priority-based routing. All strategies include low-latency fallback mechanisms to ensure resiliency

Support Popular LLM providers

Centralized configuration and API key management

Always know what’s going on with the heart of your system via logs, metrics, and traces. OpenTelemetry is natively supported (coming soon).