Unlimited text-to-speech for everything with a voice.
Drops into what you already run, agents over WebSocket, narration and dubbing over HTTP
LiveKit
Pipecat
Vapi
Retell
Daily
WhisperThe console
Type a line. Hear it back.
The engine
The engine, played back.
One identity · 23 languages
The same voice, four languages.
Ten seconds in, a twin out
A reference clip of about ten seconds becomes a working clone, no training job, no per‑voice fee, cached after first use.
How cloning works →
The whole run, not just the median
First audio byte in 146 ms over the open internet. A conversation is judged by its slowest turn, not its median one, so the number to hold us to is the tail.
The full benchmark →
Three layers under every request
Redundant serving behind one endpoint.
The serving tiers →
Pressed with a seal, trained on nothing
An inaudible watermark is stamped into every clip, and nothing you send enters a training set.
The signed clip →Everything is learning to speak. Gandr exists to be its voice. Everywhere, unmetered, measured in milliseconds.
the stream
One stream.
Unlimited speech.
Speech in 23 languages, and voice cloning from about ten seconds of audio. Nothing on a stream is counted, your busiest day costs what your quietest does.
Hear the finished work.
Two takes, pulled from the shelf.
Narration
The lighthouse keeper, chapter one
A cloned voice reads long-form
The stream
The same request, streamed
First audio inside a breath
Access
Get a key.
Sign in, and your key is on the screen. From 500 lines we price it on a call instead.
Hear it on your own script.
A key, your text, your voices, the production API. Sign in, and it is issued on the spot.
