A project by Virat Kaukuntla
rat.audio
Scripts in. Narration out, in seconds.
rat.audio is a free text-to-speech studio made by Virat Kaukuntla. Paste a script, choose a voice profile, and it reads the script aloud with Kokoro-82M, an open 82-million-parameter voice model that runs right in your browser on WebGPU. There is no server and no API key, and your script never leaves your device. It is 100% free, forever: the first generation needs no account, and after that a free Google sign-in keeps it unlimited. Audio starts playing after the first sentence, and the full track downloads as a WAV file. It is in public beta.
What it does
- Five studio voices, including Deep Male Narrator and Rocco Spinelli, a swaggering cowboy
- A Voice Lab to mix up to three of Kokoro's 28 voices, tune pitch and pace, save them and share them by link
- Record your voice (beta): it measures your pitch, tone and pace on your device and builds the closest matching voice
- A free connector (MCP) so ChatGPT and Claude can narrate scripts and design voices for you
- Runs on WebGPU with a WebAssembly fallback, at roughly an eighth of a second per word on a recent Mac
- Streams each finished sentence to the browser so playback starts before the whole script is done
- Joins segments with natural pauses for sentences and paragraphs, and exports a 24 kHz WAV
- Live pipeline console, waveform view, word counter and adjustable pace
Made by Virat Kaukuntla (Virat Reddy Kaukuntla), a high school software developer in Texas who also built Mnemo, VRK, Chroni, Animedle, Strata, AniRadar, Skelr, Pong Rivals, re://battlegrounds.
Last updated