Building with large language models means standing on a stack of open-source frameworks. The list below ranks 20 of the most popular open-source AI development frameworks and LLMOps tools by GitHub stars (fetched 2026-08-17) — from low-level inference engines to full visual app builders. Follow each link to read the source, license and release activity.
Pick by job: use an inference engine (vLLM, llama.cpp) to serve models, an orchestration library (LangChain, LlamaIndex, Haystack) to wire retrieval and tools, and an LLMOps platform (Dify, Flowise, Open WebUI) to ship a product.
