any-llm

Why we included this project

Python teams that have built their service against one model provider often discover switching is painful. any-llm offers a single completion call that can talk to OpenAI, Anthropic, Azure, Mistral, Ollama, and others, so moving providers means changing a provider name and setting the right environment key instead of rewriting request code. It deliberately follows LiteLLM's conventions, so API keys and environment variables from an existing setup carry over unchanged and migration stays cheap. The library is for developers embedding model calls into their own applications, not for end users chatting with a bot. If you are early in a build and want to keep provider options open without committing to per-provider client code, this gives you a thin layer to build on and swap out later.