awesome-llm-json

Why we included this project

Getting an LLM to return valid, schema-shaped JSON is a recurring practical headache when you build agents and pipelines, and this list gathers the scattered solutions in one place. It groups hosted and local models, Python libraries, notebooks, and blog posts by the techniques people actually use, from function calling and JSON mode to tool use and grammar-constrained generation. The terminology section alone is worth the visit, since the same capability goes by different names across providers. Per-provider tables link straight to API docs and pricing, which helps when you are comparing structured-output support, and the local-model and library sections narrow down options for self-hosted setups. It is a reference index rather than something you run, but as a starting point for comparing approaches and vendors it saves real research time.