Voice I/O API

Give your AI agent a voice and ears. Built for AI agents and creators. Turn text into studio-quality speech, transcribe any audio into text, and browse the available voices and models on demand. Pay per call from your workflow wallet (~$0.01 per synth or transcript), only when your agent runs on live audio. • 3 tools: text-to-speech, speech-to-text, models list • Under the hood: Deepgram, OpenAI • Works in Claude Code, Cursor, Codex, and any MCP client • Free demo mode always-on: every tool returns a cached PII-reduced preview with no upstream payment. Pass `demo: false` and pay via x402 to unlock live data.

Base#45322LiveAnalyticsx402 declaredMCPA2AAPIweb