OpenRouter shows how to build a voice AI agent in Python
AIOpenRouter's tutorial builds a restaurant booking voice agent that chains speech-to-text, an LLM with tool calls, and text-to-speech through one API key. In its October 8 tests, the median server-side time to first audio was 1.9 seconds across 12 turns, and text-to-speech made up about 90% of the measured cost.
AIWhy it matters.
The tutorial shows a complete voice agent loop with tool calls, sentence-level streaming to TTS, measured latency, and cost figures that readers can compare against their own builds.
Source details
OpenRouter Blog · openrouter.aiPublished · added here
Original title: How to Build a Voice AI Agent