Why this exists
LLM requests moved from demos into production paths: support replies, agent tasks, background jobs, and internal tools. Direct API calls are still useful, but they do not give every team durable retries, idempotency, and request status by default.
ReqRun exists to make that reliability layer simple, OpenAI-compatible, and focused.
Product boundary
ReqRun is not a provider marketplace, analytics platform, workflow builder, or multi-provider router. The v1 product focuses on reliable OpenAI-compatible request execution.