OpenRouter / Text to speech

Turns text into speech with one of 30 voice models on OpenRouter: ElevenLabs (v4, v3, flash, turbo, multilingual), Fish Audio (s2.1-pro, s2-pro, drama-3), MiniMax speech 2.8, Deepgram Aura 2 and Flux, Microsoft MAI Voice, Qwen TTS, Grok voice, HeyGen voice, Voxtral, Kokoro, Orpheus. Up to 5,000 characters. Returns an audio file (mp3, or raw 16-bit pcm). voice is model-specific (for example george or sarah on ElevenLabs, aura-2-thalia-en on Deepgram, af_heart on Kokoro); openrouter-models with output_modalities speech lists them per model. instructions sets tone on models that take it. Charged per character of text at the model's OpenRouter price plus its 5.5% credit fee; a refused call costs nothing.