# MWT Voice – A Free AI Text-to-Speech Studio with Kitten TTS, Kokoro TTS & Pocket TTS > Source: > Published: 2026-11-01 16:23:08+00:00 # I’m excited to introduce **MWT Voice**, a browser-based AI voice generation and audio production studio built by MWT Labs. MWT Voice brings three AI text-to-speech engines together in one application, alongside voice cloning, multi-speaker podcast generation, waveform-based audio editing, audio merging, and media conversion tools. What makes MWT Voice particularly interesting is that it is designed to work as a **fully offline-capable Progressive Web App (PWA)**. After the initial setup and model downloads, users can install it on their devices and continue using its supported features without an internet connection. No cloud-based TTS inference, no subscriptions, and no recurring AI credits. ## MWT Voice currently integrates three different AI voice engines: ### - Lightweight English text-to-speech engine. - 8 available English voices. - Fast and lightweight voice generation. - Suitable for mobile devices and lower-powered computers. ### - 54 available voices. - Supports 8 languages. - Includes American and British English, Spanish, French, Italian, Brazilian Portuguese, Hindi, Mandarin Chinese, and Japanese. - Useful for multilingual narration and voice generation. ### - 8 preset English voices. - Custom voice cloning from microphone recordings or uploaded audio. - Create and save custom voices for reuse. - Useful for personalized voice generation. ## One of the key features of MWT Voice is its dedicated Podcast & Dialogue Studio. Users can create complete conversations involving multiple speakers instead of generating just a single voice. Features include: - Add multiple speakers or characters. - Assign different voices to different speakers. - Customize speaking speed for each speaker. - Write and edit individual dialogue lines. - Assign each dialogue line to a specific speaker. - Create interviews, discussions, dialogues, and podcast-style conversations. - Control pauses between speakers. - Generate individual dialogue audio clips. - Combine the complete conversation into one continuous audio track. - Export the final conversation as MP3 or WAV. For example, one speaker can act as a podcast host and ask questions while another responds as a guest. Users can write their own scripts and turn them into complete multi-speaker audio conversations. ## With Pocket TTS, users can: - Record a voice using their microphone. - Upload an existing audio recording. - Create a custom voice from a reference recording. - Preview the cloned voice. - Save custom voices for later use. ## MWT Voice also includes an audio editing workspace with: - Waveform-based audio editing. - Audio trimming. - Split audio at the playhead. - Peak normalization. - Reverse audio. - Fade-in and fade-out. - Pitch adjustment. - Formant shifting. - Voice changing and audio tone adjustments. Users can send generated speech directly into the Audio Studio for further editing. ## Combine multiple audio clips into a single track. - Add multiple audio files. - Reorder clips. - Configure custom silence gaps. - Preview individual clips. - Export combined audio. Useful for assembling podcast episodes, interviews, voiceovers, and other audio projects. ## Built-in conversion tools support: - WAV to MP3 conversion. - MP3 to WAV conversion. - MP3 export up to 320 kbps. - Extract audio from MP4 videos. - Extract audio from WebM videos. ## **MWT Voice is designed to work like an installed desktop or mobile application, not just a website that requires a permanent internet connection.** Its offline capabilities are one of its core advantages. Here’s how it works: 1. Open MWT Voice while connected to the internet. 2. Download the required AI models directly to your device. 3. Install MWT Voice as a Progressive Web App (PWA). 4. Allow the application and its required resources to be cached. 5. Once setup is complete, open the installed application without an internet connection. After the initial setup, supported features can continue working offline: - Generate AI speech using downloaded TTS models. - Use supported voice cloning functionality. - Create multi-speaker podcasts and conversations. - Edit generated audio. - Merge multiple audio files. - Convert supported audio and video formats. - Access the installed application without repeatedly visiting the website. ### - AI speech generation runs locally on the user’s device. - No cloud-based TTS inference. - No account registration required. - No subscription or recurring AI credits. - Text and generated speech are processed locally. - Downloaded models are stored on the device. - Users can manage and delete downloaded models. - Installable PWA experience for desktop and mobile devices. **Important:** An internet connection is needed initially to access/install the application and download models that have not yet been cached. Once the required models and application resources are available locally, supported workflows can operate without internet access. Actual performance depends on the device’s processing power, available memory, browser capabilities, and selected AI model. ## MWT Voice can be useful for: - YouTube creators and video producers. - Podcast creators. - Content creators. - Students and educators. - Audiobook and storytelling projects. - Developers experimenting with browser-based AI and local inference. - Anyone who needs text-to-speech or everyday audio editing tools. ## **MWT Voice:** [https://voice.mwtlabs.uk](https://voice.mwtlabs.uk) **Developer:** MWT Labs **Website:** [https://mwtlabs.uk](https://mwtlabs.uk) MWT Voice is free to use and supported by non-intrusive advertisements. I’d love to hear feedback from the Hugging Face community, especially from people interested in browser-based AI, local inference, offline AI applications, text-to-speech, voice cloning, and multi-speaker audio generation. If you try it, let me know what you think!