Calculate and optimize LLM response latency — TTFT, TPS and end-to-end latency metrics. Built for AI developers.