Tempr Gateway

An OpenAI-compatible API in front of the provider keys you already have. Swap your SDK base URL and get virtual keys (model allowlist, budget, RPM/TPM limits), searchable request logs, and analytics by key and model. No token markup: you pay your provider directly and Tempr by request volume. Plus fallback chains, prompt caching, custom models on your own vLLM/SGLang servers, and the Responses API on any model. Free tier: 10,000 requests/month, no card.