The Render demo intentionally runs the Streamlit application as a single web service:
Streamlit -> OpenAI-compatible HTTPS endpoint -> LLM
The local LiteLLM/Ollama Compose files remain for development and private/local inference. They are not required for the public Render demo.
Set OPENAI_API_KEY in Render. The example render.yaml uses OpenRouter's OpenAI-compatible API endpoint, but any compatible endpoint can be used by changing OPENAI_BASE_URL and OPENAI_MODEL.
- Runtime: Docker
- Region: Singapore
- Health check:
/_stcore/health - Start command: provided by the Dockerfile
- Persistent disk: not required for the demo
The Free deployment is a portfolio/demo service, not a production procurement system. Files are processed in temporary storage during a request and are not intended to become a durable audit repository.