Quick Answer
Paste text, pick a voice, and press Play — your browser synthesizes the speech on your device in real time. Browser voices never send your text anywhere; optional neural voices sound more natural and can download as MP3. The local AI voice option (Kokoro) generates natural speech entirely in your browser: a one-time ~90 MB model download, ten English voices, speed control, and WAV export — free, no signup.
How It Works
Enter or paste your text into the input field
Select your preferred voice from the available options
Adjust speed and pitch settings if desired
Click Play to preview or Download to save as MP3
For natural local AI speech, switch to Local AI voice (Kokoro), let the model download once, then play or export as WAV
Key Facts
- Convert any text to natural-sounding speech instantly
- Choose from multiple system voices or AI neural voices
- Adjust speech rate and pitch to your preference
- Download audio as MP3 for offline use
- All processing happens in your browser - fully private
- Supports multiple languages and accents
- Complete privacy — browser voices never leave your device
- No data logging or retention for any voice option
- Local AI voice (Kokoro): ~82M-parameter model runs in your browser via WebGPU or WebAssembly — one-time ~90 MB download, ten English voices, WAV export, nothing uploaded
Frequently Asked Questions
Is text to speech free to use?
Yes, this text to speech tool is completely free with no hidden costs. There are no limits on usage, no signup required, and no watermarks on the audio output.
Is my text private?
Absolutely. All text processing happens locally in your browser. Your text is never sent to any server. This makes it safe for sensitive, confidential, or personal content.
What's the difference between Standard and Neural voices?
Standard mode uses your device's built-in voices for instant playback with zero download. Neural mode uses AI-powered Piper voices that sound more natural and support MP3 download, but require a one-time voice model download (15-25MB per voice).
Can I download the audio as MP3?
Yes, but only with Neural mode. Switch to Neural, download a voice, and then use the Download MP3 button. Standard mode uses the Web Speech API which does not provide access to the raw audio data.
What browsers are supported?
The tool works on all modern browsers including Chrome, Firefox, Safari, and Edge. Standard mode requires browser speech synthesis support. Neural mode requires WebAssembly support, which all modern browsers have.
How do neural voice downloads work?
When you switch to Neural mode and select a voice, you can download the voice model (15-25MB). The model is stored in your browser for offline use. You only need to download each voice once.
What languages are supported?
Standard mode supports all voices installed on your system, which varies by device. Neural mode currently includes US English, British English, German, Spanish, and French voices, with more being added.
Is my text used for AI training?
No. Browser voices process entirely locally. AI voice providers do not retain or use your text for training. The local Kokoro engine generates speech on your own device, so nothing is uploaded at all.
What is the local AI voice option?
The Local AI voice engine uses Kokoro, an open-source text-to-speech model with roughly 82 million parameters that runs directly in your browser using WebGPU (or WebAssembly on devices without it). The model files (~90 MB) download once from a CDN and are cached by your browser. After that, speech is generated entirely on your device — your text is never uploaded anywhere.
Is the local AI voice really free?
Yes. Kokoro is open-source and runs on your own hardware, so there are no per-character charges, subscriptions, or usage caps — nothing to bill. Commercial neural voice services such as ElevenLabs meter usage and bill monthly once you outgrow their limited free tier; this engine costs nothing beyond the one-time model download.