llmflows

Why we included this project

LLMFlows is a lightweight Python framework for building multi-step LLM applications without inheriting a heavyweight orchestration layer. Its core is a small set of primitives: LLM wrappers, prompt templates, and Flow/FlowStep classes that let you chain dependent calls, run independent steps in parallel, and keep every prompt and model call visible in code. That explicit, no-hidden-prompts design makes it easy to debug exactly what your app sends to the model. It also ships chat abstractions with message history and vector-store integration, so building a chatbot or question-answering pipeline stays straightforward. If you want a framework you can read end to end rather than a large platform with many moving parts, this is worth a look for your next LLM project.