Try Kokoro TTS Online for Free
One of the smallest models with natural-sounding voices. Runs entirely in your browser. No signup, no install, completely free.
Try Kokoro FreeWhy Kokoro
Why Use Kokoro for Text-to-Speech
Kokoro is an open-source, Apache 2.0-licensed text-to-speech model with just 82 million parameters. Built on the StyleTTS 2 architecture by hexgrad, it delivers speech quality that consistently outperforms models many times its size, including XTTS at 467M parameters and MetaVoice at 1.2B parameters.
At roughly 80 MB, Kokoro is small enough to run directly in your browser while still producing natural, expressive speech across 8 languages.
27+ Built-in Voices
Choose from a diverse collection of voices including af_heart, af_bella, af_sarah, am_adam, and many more across different languages and styles.
8 Languages
Generate speech in English (American and British), Japanese, Chinese, Spanish, French, Hindi, and Italian.
Ultra-Lightweight
Only 82 million parameters and roughly 80 MB in size. Runs fast on almost any device without a dedicated GPU.
Speed Control
Adjust playback speed from 0.5x to 2.0x to match your content needs, from slow narration to fast-paced dialogue.
Get Started
How It Works
Open the free tool
Go to the Kokoro TTS tool in your browser. No download or signup required.
Pick a voice
Browse the 27+ built-in voices and select one that fits your project. Each voice has a prefix indicating its language and gender, so you can quickly find the right match.
Type your text and generate
Enter the text you want spoken, adjust the speed if needed, then hit generate. Your speech is ready in seconds.
Use Cases
Who Is Kokoro For?
Content Creators
Add natural voiceovers to YouTube videos, podcasts, or social media content. Choose from 27+ voices across 8 languages without recording anything yourself.
Educators
Create audio versions of lessons, study guides, and learning materials. Use different voices for different characters or topics to keep students engaged.
Developers
Prototype voice features, test TTS pipelines, or integrate speech output into applications. Kokoro is Apache 2.0 licensed, so you can use it freely in your projects.
Accessibility
Convert written content to spoken audio for people who prefer or need to listen rather than read. Supports multiple languages for a wider audience.
Getting the Most Out of Kokoro
Tips for Best Results
Choose the right voice prefix for your language
Voice names start with a prefix that indicates language and gender. For example, af_ means American female, am_ means American male, and bf_ means British female. Matching the voice to your text language gives the best pronunciation.
Adjust speed for your context
Use slower speeds (0.5x to 0.8x) for narration or educational content where clarity matters. Use faster speeds (1.2x to 2.0x) for quick previews or dialogue-heavy scenes.
Try different voices for different tones
Each built-in voice has its own character and tone. Experiment with a few voices to find the one that best matches the mood of your content, whether it is warm, professional, or energetic.
Voice Creator Pro
Need more? Go further with Voice Creator Pro.
Voice Creator Pro includes GPU acceleration, voice cloning from short audio samples, voice design from text descriptions, TTS across 600+ languages, a local REST API, and a commercial use license. One-time purchase of $49.99.
Voice Cloning
Clone any voice from a short audio sample. Zero-shot cloning with no training required
600+ Languages
Combine multiple open-source models for TTS across 600+ languages
GPU-Accelerated Processing
Faster generation with NVIDIA, Apple Silicon, AMD, and Intel GPU support
Voice Design from Text
Describe a voice in plain text and the AI creates it. No audio samples needed
Local REST API
Automate voice generation in your own apps and workflows
Commercial License
Full rights to use generated audio in commercial projects. One-time $49.99 purchase
FAQ
Common Questions
Kokoro is an open-source text-to-speech model created by hexgrad. It has 82 million parameters and is built on the StyleTTS 2 architecture. Despite its small size, it produces natural-sounding speech that rivals models many times larger.
Kokoro was developed by hexgrad, an open-source community contributor. It is released under the Apache 2.0 license, which means anyone can use, modify, and distribute it freely, including for commercial purposes.
Kokoro supports 8 languages: English (American and British), Japanese, Chinese, Spanish, French, Hindi, and Italian. Each language has dedicated voices optimized for natural pronunciation.
Kokoro includes 27+ built-in voices. Each voice has a prefix that indicates its characteristics. For example, af_heart, af_bella, and af_sarah are American female voices, while am_adam is an American male voice. You can browse and preview all available voices in the tool.
Yes. Kokoro is open-source software released under the Apache 2.0 license, one of the most permissive open-source licenses available. On this site, it runs directly in your browser with no account, no signup, and no usage limits.
Kokoro consistently surpasses much larger models in speech quality benchmarks despite having only 82 million parameters. It outperforms XTTS (467M parameters) and MetaVoice (1.2B parameters) while being significantly faster and smaller. Its lightweight design means it can run efficiently in a browser without needing a powerful GPU. For even more capabilities, including voice cloning and 600+ languages, Voice Creator Pro offers a full desktop solution.
No. Kokoro is only about 80 MB in size and has just 82 million parameters, making it one of the lightest TTS models available. It runs well on most modern computers and does not require a dedicated GPU. Any recent laptop or desktop should handle it without issues.
Chrome, Edge, and other Chromium-based browsers work best. Firefox and Safari have limited support for the WebGPU features that Kokoro uses for acceleration.