Overview
Call 100+ LLM providers with OpenAI-compatible API. Budget tracking and fallbacks built-in.
Essential glue layer when your stack uses more than one LLM provider.
Key features
- Single OpenAI-compatible endpoint for 100+ providers
- Budget caps, rate limits, and fallback routing
- Spend tracking per API key and team
- Drop-in replacement for OpenAI SDK calls
Use cases
- 01Multi-model routing (GPT-4o → Claude → local Ollama)
- 02Centralized LLM gateway for microservices
- 03Cost control for AI features in production SaaS
Quick start
pip — quick start
pip install litellm
litellm --model gpt-4o-miniStack compatibility
Ollama
Open WebUI
LangChain
Vercel AI SDK
Tags
Related articles
AI agents for customer support: CrewAI vs AutoGen
Compare multi-agent frameworks for automating tier-1 support in indie SaaS.
Best open-source LLM tools for indie hackers (2026)
Ollama, vLLM, LiteLLM, and more — what to self-host when you're bootstrapping AI features.
Fine-tuning Llama 3 on a budget GPU
LoRA fine-tuning on RunPod for under $20 — when it makes sense vs prompt engineering.