Stay free if you only need 100+ llm provider integrations and langfuse, arize phoenix, langsmith, otel logging. Upgrade if you need everything in open source and jwt authentication and sso integration. Most solo builders can start free.
Why it matters: Requires Docker and infrastructure knowledge for self-hosted deployment
Available from: Enterprise
Why it matters: Enterprise features like SSO and audit logging locked behind paid tier
Available from: Enterprise
Why it matters: Enterprise pricing requires sales consultation with no published rates
Available from: Enterprise
Why it matters: Configuration complexity increases significantly with many providers and routing rules
Available from: Enterprise
Why it matters: Limited built-in UI for non-technical users — primarily CLI and API-driven
Available from: Enterprise
Why it matters: Observability integrations require separate setup of Langfuse, Grafana, etc.
Available from: Enterprise
Yes. LiteLLM is available as a Python package (pip install litellm) that you can use as a library in your code or run as a standalone proxy server. Docker is recommended for production deployments but not required.
LiteLLM adds minimal overhead — typically under 10ms per request for local proxy deployments. The proxy handles routing, logging, and spend calculation asynchronously to minimize impact on response times.
Direct provider SDKs lock you into a single provider. LiteLLM gives you automatic failover across providers, unified spend tracking, budget enforcement, and the ability to switch models by changing a parameter — without rewriting application code.
LiteLLM's self-hosted proxy runs entirely on your infrastructure. No data passes through LiteLLM's servers. For the enterprise cloud option, LiteLLM provides security documentation and compliance FAQs at docs.litellm.ai/docs/data_security.
LiteLLM supports 100+ providers including OpenAI, Anthropic Claude, Google Gemini, AWS Bedrock, Azure OpenAI, Cohere, Mistral, Together AI, Replicate, Hugging Face, Ollama for local models, and many more. New providers are added regularly.
Yes. LiteLLM supports routing to local model servers including Ollama, vLLM, and any OpenAI-compatible endpoint. This allows you to mix cloud and local models in the same routing configuration with unified logging and spend tracking.
Start with the free plan — upgrade when you need more.
Get Started Free →Still not sure? Read our full verdict →
Last verified March 2026