LLM fact extractor for any OpenAI-compatible chat client, conforming to the spec. Works with
OpenAI itself and with local or self-hosted servers that speak the same API (Ollama, vLLM, LM
Studio): pass an openai client built with their baseURL. It asks for a JSON-schema structured
output and, if a server rejects that, falls back to plain JSON mode for the rest of its life.
LLM fact extractor for any OpenAI-compatible chat client, conforming to the spec. Works with OpenAI itself and with local or self-hosted servers that speak the same API (Ollama, vLLM, LM Studio): pass an
openaiclient built with theirbaseURL. It asks for a JSON-schema structured output and, if a server rejects that, falls back to plain JSON mode for the rest of its life.