
LiteLLM is an open-source AI gateway company serving platform teams across model providers.
LiteLLM's offering centers on an open-source AI gateway for platform teams. The gateway exposes large language model APIs from providers such as OpenAI, Anthropic, Microsoft Azure, Amazon Bedrock, Google Vertex AI, and Cohere through one OpenAI-compatible interface, used through a Python SDK or a self-hosted server.
Around the gateway the company runs a commercial enterprise tier with administration, security, and support capabilities, plus a managed agents platform for operating multiple agents. The product family targets engineering teams that need routing, load balancing, fallbacks, and spend management across model providers.
LiteLLM's gateway is open source under an MIT license, which lets engineering teams self-host, audit, and extend the software rather than depend on a closed managed service. Supporting many model providers behind one OpenAI-compatible interface reduces the work of switching or combining vendors.
The company monetizes an enterprise tier built on the same open-source core rather than a separate codebase, so adoption of the free gateway and the paid offering reinforce each other. Enterprise capabilities include single sign-on, audit logs, and service-level agreements.
LiteLLM distributes its core gateway as free, MIT-licensed open source, including the Python SDK and the self-hosted proxy server. Paid value is concentrated in the enterprise tier, sold as a commercial license with security and administration features such as single sign-on, audit logs, and service-level agreements.
This open-core posture lets self-hosting teams adopt the gateway at no cost while organizations with compliance and support needs pay for the enterprise layer, which is also procurable through cloud marketplaces. The company additionally offers managed deployment options for teams that prefer not to operate the gateway themselves.