OfflineTTS is Free
The current OfflineTTS browser application has no subscription, signup, API key, or per-character charge. Your device supplies the storage and compute, and third-party services linked from this site are separate.
Text to Speech
- ✅ 97 voice options and presets across 32 TTS languages
- ✅ 4 engines: Kokoro, Kitten, Piper, Supertonic
- ✅ No OfflineTTS per-character meter
- ✅ WAV and MP3 export
- ✅ Local synthesis after required assets load
- ✅ No API key needed
Speech to Text
- ✅ 99 languages supported
- ✅ Optional precise word-level timestamps
- ✅ 3 model sizes (Tiny/Base/Small)
- ✅ TXT, SRT, VTT export
- ✅ Streaming transcription
- ✅ Local transcription after model download
What “Free” Includes
You can open the public OfflineTTS tools and use the included TTS engines, Whisper transcription, document reader, text-preparation utilities, playback, and available export controls without buying an OfflineTTS plan. We do not ask for payment details, issue a monthly quota, or charge for each character processed in the browser.
“Free” describes the current OfflineTTS product price. It is not a promise that every upstream model, optional outside service, internet connection, device, or future product will always have no cost. A paid service reached through an affiliate or comparison link is provided and billed by that company under its own plans.
How Can It Be Free?
OfflineTTS runs AI models directly on your device using WebGPU or WebAssembly. There are no server costs for speech generation or transcription because your hardware does the work. The browser app is static, and the AI work runs locally on your hardware, so operating costs stay low. Some network infrastructure is still used for site delivery, model downloads, analytics, and non-English Kokoro phonemization. Keeping the expensive speech inference on the device makes the free product practical.
Costs Your Device Still Handles
Local processing shifts practical resource costs to your browser and hardware. Before choosing a larger model or a long workflow, account for:
- internet bandwidth for the application and initial model download;
- browser storage for cached model files and space for exported audio or subtitle files;
- CPU or GPU time, memory, battery use, and electricity during generation or transcription;
- another download if browser storage is cleared, the model changes, or you switch profiles or devices; and
- a network request containing text when non-English Kokoro needs server-side phonemization.
“Offline” is engine- and language-specific. The first load needs a connection, and non-English Kokoro is not a fully disconnected text workflow. Check the Privacy Policy for the exact data path.
Commercial Use and Upstream Terms
OfflineTTS does not claim ownership of the audio file you generate and does not sell a separate commercial license for the browser application output. That does not automatically clear every use. The applicable upstream model or voice terms, rights in your source text or recording, personality and trademark rights, and the rules of the platform where you publish may still apply.
Review the exact model and voice documentation before commercial distribution, especially when using Piper voices assembled from different datasets. The model-source table on the Transparency page is a starting point, not legal advice and not a substitute for checking the current upstream release.
No OfflineTTS Usage Charge
No Subscriptions
No monthly or annual plans
No Per-Character Fees
No browser-tool character meter
No API Keys
Just open and use