rust-genai

Why we included this project

Rust developers who want one code path for talking to several LLM providers without adopting each vendor's SDK will find this crate useful. The API works directly with Anthropic, OpenAI, Gemini, and local Ollama instances, and model names can carry a namespace like qwen_cloud::qwen3.8-max so you switch providers by editing a string rather than rewriting calls. It also supports custom OpenAI-compatible endpoints through environment variables or a target resolver, which fits teams that run self-hosted models behind a compatible service. Because requests go through each provider's native protocol instead of a translation layer, you keep provider-specific behavior while having a single interface in your codebase. That makes it a practical option for small teams looking to standardize LLM access across many models.