Gandr

Gandr is a text-to-speech platform that turns written text into spoken audio. It's built for developers who need to add voice to their apps, from AI agents to audiobooks. Unlike services that charge per character, Gandr uses a flat per-stream pricing model — you pay for the stream, not the words. It supports 23 languages and can clone a voice from just ten seconds of audio, no training job required.

The platform works over WebSocket for real-time agent applications and over HTTP for narration and dubbing. It integrates with popular frameworks like LiveKit, Pipecat, Vapi, Retell, and Daily. The engine is fast: first audio byte arrives in about 146 milliseconds over the open internet, and the 95th percentile latency is still under 200 ms, making it suitable for conversational use. Every clip is watermarked with an inaudible provenance seal, and Gandr does not train on your data. Voice cloning is cached after the first use, so there's no per-voice fee.

Gandr is best for developers building voice AI agents, content creators producing audiobooks or multilingual dubbing, and enterprises that need reliable, low-latency TTS at scale. The flat pricing makes costs predictable, and the simple API lets you get started with a key in seconds.