connect-llm-simulator.md

Use the LLM Simulator

Don't have an endpoint yet? The included LLM Simulator returns realistic Azure OpenAI / OpenAI / Anthropic / Gemini-shaped responses — including 429 throttling and streaming — so you can validate the full proxy pipeline first.

cd test/LLMSimulator && func start

export Host1="host=http://localhost:7071;probe=/api/health"

Point your existing client code at the proxy unchanged — the simulator mirrors real provider URL shapes, so nothing else needs to change when you swap in a real endpoint later.

See → LLM Simulator for the full list of simulated models and error scenarios, and → Dummy Backend for an even simpler mock option.

---

← Back to Where are your LLM models?