About LLM Gateway

What we do

Teams building with large language models juggle multiple provider accounts, incompatible SDKs, separate invoices, and no unified view of what their AI features actually cost. LLM Gateway sits between your application and every provider: you send requests to one endpoint, and the gateway routes them to OpenAI, Anthropic, Google, Meta, Mistral, DeepSeek, and dozens of other providers — with per-request cost tracking, response caching, and automatic failover when a provider degrades.

The platform has routed 100B+ tokens across 20M+ requests, and adds new models within 48 hours of their provider release. Developers use the hosted service on pay-as-you-go credits with a flat 5% platform fee, bring their own provider keys for free, or self-host the entire stack with one Docker command.

Who we are

LLM Gateway is operated by Polar Lights LLC, 16192 Coastal Highway, Lewes, DE 19958, United States, and developed in the open by a distributed team and community of contributors on GitHub. The full source is licensed under AGPLv3, with commercial enterprise licensing available for the features in the ee/ directory.