Home / Directory / Text to speech
Alternatives to ElevenLabs
Turn text into speech on your own computer or phone, with voices you can keep.
Ranked by score out of 30, most convivial first.
Kokoro
A small open-weight text-to-speech model with Apache-licensed weights that runs on modest hardware.
eSpeak NG
A compact speech synthesizer that covers more than a hundred languages and accents.
Piper
A fast, local neural text-to-speech engine you install with one command.
SherpaTTS
An Android text-to-speech engine that uses Piper or Coqui voices and reads aloud with no connection.
Chatterbox
An MIT-licensed text-to-speech model with voice cloning, from a company that sells voice AI.
For comparison
The big names these tools are measured against. We review them as benchmarks, so they sit outside the directory.
About these tools
Speech that sounds like a person used to require sending your text to a company. Piper and Kokoro run neural voices locally, Chatterbox adds voice cloning, eSpeak NG is the tiny classic that covers more than a hundred languages, and SherpaTTS brings Piper voices to Android.
Most of these are building blocks for people who can run a command line, so expect some setup, and listen to samples before you commit, since voices differ more than the licences do.
How to choose
Decide where you need the voice. On a phone, SherpaTTS. In a script or a home server, Piper or Kokoro. For cloning, Chatterbox, knowing its output is watermarked. For the widest language coverage on almost any device, eSpeak NG.
Frequently asked questions
Is there a free alternative to ElevenLabs?
Piper, Kokoro and Chatterbox are free and run on your own computer, and SherpaTTS does the job on Android.
Can I clone a voice locally?
Chatterbox says it clones from a short reference clip, and every file it creates carries a Resemble AI watermark.
Which one covers the most languages?
eSpeak NG supports more than 100 languages and accents, though its voices sound more robotic than neural ones.