llm-chain

Why we included this project

Rust teams building LLM-backed services get a structured way to compose multi-step prompt pipelines here, instead of hand-wiring every API call. You define reusable prompt templates and chain them so a task no single model call can handle gets broken into ordered steps, with support for cloud providers like OpenAI and locally hosted models such as LLaMa and Alpaca. The crate also ships vector store integrations for long-term memory, plus a tools layer that lets agents run shell commands, execute Python, or search the web mid-chain. That combination is a practical base for chatbots and summarizers, and it extends to agent-style applications in a systems language. For teams already committed to Rust, the typed executor and prompt macros are a more ergonomic fit than wrapping a Python framework from their codebase.