Frequently Asked Questions
Everything you need to know about OfflineTTS — free AI text to speech and speech to text tools.
General
How does OfflineTTS work?
OfflineTTS runs AI models directly in your browser using WebGPU or WebAssembly. Choose from 4 TTS engines and use Whisper STT across 99 languages.
Is it really free?
The current OfflineTTS browser application is free to use: no subscription, account, API key, or per-character charge. OfflineTTS includes Kokoro, Kitten, Piper, and Supertonic for TTS plus Whisper for STT. Your device still supplies download bandwidth, storage, compute, and battery power, and third-party services linked from the site have their own prices.
Is the OfflineTTS application open source?
No. The OfflineTTS application, site integration, design, and branding are proprietary. It incorporates upstream models and runtimes such as Kokoro, Kitten, Piper, Supertonic, Whisper, and ONNX Runtime Web, whose own licenses continue to apply to those components. An upstream open-source license does not make the OfflineTTS application source public.
Does it work offline?
Offline behavior depends on the selected engine and language. English Kokoro works locally after its 305MB q4 or 326MB fp32 model is cached. Kitten, Piper, and Supertonic also synthesize locally after their model files download. Non-English Kokoro uses the OfflineTTS phonemization service before audio is synthesized on your device. Whisper STT runs locally after its model download.
What browsers are supported?
Recent desktop Chrome and Edge releases generally provide the most reliable WebGPU path. Safari and mobile support depends on the operating-system version, device, memory limits, and the selected engine. A WebAssembly fallback is available for supported workflows, but support and performance are not identical in every modern browser. If a model fails, include the exact browser version and backend in a bug report.
What is text to speech (TTS)?
Text to speech (TTS) is technology that converts written text into spoken audio. Modern TTS uses neural network models to generate natural-sounding human speech. Unlike older robotic-sounding systems, neural TTS models like Kokoro, Kitten, Piper, and Supertonic produce speech that sounds closer to a real person reading your text aloud.
What is speech to text (STT)?
Speech to text (STT), also called speech recognition or transcription, converts spoken audio into written text. OfflineTTS uses OpenAI's Whisper model, which supports 99 languages. Fast mode provides segment timing, while the optional Precise subtitles mode adds word-level timestamps for synchronized playback, subtitles, meeting notes, and searchable audio.
Voices & Languages
What voices are available?
97 voice options and presets across 32 TTS languages, plus 99 languages for speech to text. Choose Kokoro for its broad voice library, Piper for CPU speed, Kitten for a lightweight model, or Supertonic for local multilingual generation.
What languages are supported for TTS?
TTS covers 32 languages. Supertonic 3 provides 31 languages, while Kokoro adds Mandarin Chinese plus regional American and British English voice sets. Kokoro also provides voices for Japanese, Spanish, French, Hindi, Italian, and Brazilian Portuguese.
What languages are supported for STT?
Whisper's multilingual speech-to-text model supports 99 languages, including English, Spanish, French, German, Chinese, Japanese, Korean, Arabic, Hindi, Portuguese, Russian, Italian, and Dutch. OfflineTTS provides a selector for common languages so you can match the model to the recording before transcription.
Which voice should I use?
For English: Heart (A-rated, warm storytelling), Bella (A-rated, energetic vlogs), Michael (B-rated, professional reviews). For other languages, each has curated voices optimized for natural pronunciation. Try different voices to find the one that matches your content style.
What are voice quality ratings?
The A–D labels shown for some Kokoro voices are internal catalog guidance, not a standardized benchmark or a guarantee for every script. They summarize editorial impressions of naturalness, pacing, pronunciation, and audible artifacts on sample text. Language, punctuation, names, numbers, and speaking style can change the result, so listen to the same representative passage in several voices before choosing one.
What are Kitten TTS expression voices?
Kitten TTS uses 8 expression embeddings instead of individual voice models: cheerful, serious, sad, whisper, excited, gentle, calm, and neutral. Each expression shapes the tone and delivery style of the output. This approach gives you creative control over the emotional character of the speech while keeping the model extremely lightweight (24MB).
Privacy & Data
Is my text data safe?
English Kokoro does not send your text to the phonemization endpoint after required assets are available. For non-English Kokoro (Japanese, Chinese, Spanish, French, Hindi, Italian, Portuguese), your text is sent to api.offlinetts.com and returned as pronunciation data before local synthesis. The application is configured not to retain submitted text as content. Audio synthesis and Whisper transcription inference run on your device; normal website, model-delivery, and analytics requests are documented in the Privacy Policy.
What data do you collect?
OfflineTTS does not require an account and the browser application is not designed to upload English TTS text, STT recordings, or generated audio for inference. Non-English Kokoro sends entered text to the documented phonemization endpoint. The website uses GA4, Microsoft Clarity, hosting, and model-delivery services that process ordinary site usage and request data under their policies. See the Privacy Policy for browser storage, cookies, deletion, and opt-out details.
Can I use generated speech commercially?
In most creator workflows, yes: you can download and use generated audio in videos, podcasts, audiobooks, and commercial projects. Kokoro and Piper use permissive upstream licenses; Kitten and Supertonic are also available as local TTS engines, but you should check the upstream model terms for the exact engine you use before large-scale commercial deployment.
What is phonemization and why does it need a server?
Phonemization converts written text into pronunciation tokens that the TTS model uses to generate speech. Kokoro's browser integration handles English locally. For Japanese, Chinese, Spanish, French, Hindi, Italian, and Portuguese, OfflineTTS sends the entered text to api.offlinetts.com, receives the phoneme result, and then synthesizes audio in the browser. The application is configured not to persist submitted text as content; normal infrastructure security and request metadata may still be processed.
Technical
What audio formats can I export?
You can export audio as WAV (lossless, studio-quality) or MP3 (compressed, smaller file size). WAV is recommended for further audio editing; MP3 is great for direct use in videos and podcasts.
How much text can I convert at once?
Up to 50,000 characters per session. Longer texts are automatically split into chunks and processed sequentially with natural pauses between segments.
Why is WebGPU recommended?
WebGPU can use a compatible GPU for browser inference and is usually the preferred path for larger models. Actual speed relative to WebAssembly depends on the engine, model, browser, drivers, and hardware, so OfflineTTS does not promise one fixed multiplier. When a supported engine cannot use WebGPU, the application may offer or select a WebAssembly path, which can be slower and use resources differently.
What model sizes are available?
Kokoro TTS offers q4 (~305MB, recommended) and fp32 (~326MB, full precision). A usable q8 build is not available for this model. Kitten is the lightest TTS option at ~24MB, Piper is CPU-oriented at ~75MB, and Supertonic loads a multi-file ONNX model stack for multilingual synthesis. For STT, approximate browser downloads are Tiny (~120MB, fastest), Base (~210MB, balanced), and Small (~590MB, best accuracy).
Does it work on mobile?
TTS and STT work best on desktop browsers with WebGPU support. Mobile browsers may have limited WebGPU/WASM support and could produce errors. For the best experience, use Chrome or Edge on a desktop or laptop.
What is the difference between Kokoro, Kitten, Piper, and Supertonic TTS engines?
Kokoro TTS has 54 voices across 9 language groups and uses WebGPU with a WASM fallback. Kitten TTS has 8 expressions in a lightweight 24MB model. Piper TTS offers 25 curated voices with CPU-friendly WASM inference. Supertonic TTS provides 10 built-in voices across 31 languages for local multilingual generation.
How are the models downloaded and cached?
Models are downloaded over HTTPS on first use and cached in your browser's IndexedDB storage. Kokoro models are served from Cloudflare R2 (CDN) and Hugging Face. Subsequent visits load instantly from cache without re-downloading. You can clear the cache through your browser's storage settings.
What is the sample rate of the generated audio?
Kokoro TTS outputs at 24kHz. Kitten TTS is configurable from 8kHz to 48kHz. Piper TTS has a fixed 22.05kHz sample rate. For STT, audio input at any common sample rate is accepted — the tool handles resampling automatically.
How does the text chunking work for long texts?
When you enter text longer than the model's optimal chunk size, it is automatically split into segments at sentence boundaries. Each chunk is processed independently and the results are concatenated with natural pauses. This ensures consistent quality even for very long texts like audiobook chapters.
Comparisons
Is OfflineTTS better than ElevenLabs?
They serve different needs. OfflineTTS focuses on free browser inference, downloadable models, and workflows that can run locally after required assets load. ElevenLabs is a hosted commercial platform with its own voice catalog, APIs, production features, plans, and data practices. Neither is universally “better”: compare a representative script, required license, current provider limits, privacy needs, editing workflow, and total cost before deciding.
How does OfflineTTS compare to NaturalReader?
OfflineTTS focuses on no-account browser speech tools and local inference where the selected engine and language support it. NaturalReader provides its own reading applications, voices, document workflows, cloud features, and plan structure. Prices and feature limits change, so verify NaturalReader's current plan page and compare document support, available voices, export rights, processing location, and the devices you need to use.
How does OfflineTTS compare to Speechify?
OfflineTTS is a browser tool centered on local speech inference and downloadable output. Speechify offers a separate commercial reading and voice platform across its supported apps and devices. Compare current pricing, voice and document features, account requirements, output rights, accessibility workflow, and privacy policy. OfflineTTS also has documented network activity for model delivery, analytics, and non-English Kokoro phonemization, so “offline” must be evaluated for the exact workflow.
How does OfflineTTS compare to browser built-in TTS?
Browser built-in TTS uses voices supplied by the operating system through the SpeechSynthesis API, so availability and sound vary by device. It can start quickly and may use little additional storage. OfflineTTS downloads neural model assets and offers a consistent in-page workflow with selectable engines and file export. Compare pronunciation, latency, storage, offline behavior, and export needs on your own device rather than assuming one voice system always sounds better.
Use Cases
Can I use OfflineTTS for YouTube videos?
Yes. Generate voice-overs for YouTube videos, download as WAV, and import into your video editor (DaVinci Resolve, Premiere Pro, Final Cut, etc.). Heart (A-rated) is the top pick for educational content, Bella for vlogs, and Michael for review videos.
Can I create audiobooks with OfflineTTS?
Yes. Import EPUB, PDF, or TXT in the Ebook to Audio tool, select chapters, and listen in the synchronized reader. For Kokoro, use q4 for the recommended quality-to-size balance or fp32 for full precision. You can export the generated audio for editing and publishing.
Can I use OfflineTTS for e-learning?
Absolutely. Add voice narration to online courses, training materials, and educational content. OfflineTTS covers 32 TTS languages across its local engines for international audiences. Generate consistent, professional narration without hiring voice talent for every course update.
Can I use OfflineTTS for accessibility?
Yes. Convert text to speech for visually impaired users, create audio versions of written content, and add voice narration to any web content. The STT tool can also generate subtitles (SRT/VTT) for making video content accessible.