Local text to speech
A local voice that runs in your browser
The voice runs on your device, not on a cloud service. Local files are read and narrated in the browser and are not uploaded. The voice is fetched once, then it can work offline.
What “local” means here
The app opens the file in your browser, keeps its structure, and speaks it a sentence at a time. The voice itself also runs in the browser. There is no cloud voice service and no upload endpoint for local files.
“Local” is about where the work happens. It does not mean the app never uses the network: the first visit loads the app and its speech runtime, and the first use of a voice fetches the voice files. A URL import is a direct request to that article.
What leaves your device
| Action | What leaves | Why |
|---|---|---|
| Open a local file | Nothing | Text and speech are local |
| First use of a voice | A one-time request for the runtime and the voice files | So speech can run on your device |
| Paste a URL | A direct request to that URL | Fetching the article |
| Show an ad | A request to the ad network, with standard request data | Serving and measuring the ad |
| Visit or install the app | Standard request metadata to the host or CDN | Serving the site and app files |
| Email the team | Your email address and message | Replying to you |
Local voice vs cloud read-aloud
| Dimension | This app | Cloud read-aloud |
|---|---|---|
| Where speech is generated | In your browser | On a remote service |
| Local file upload | Not uploaded | Usually uploaded |
| Account | Not required | Usually required |
| Works offline | After setup | Usually no |
| Languages | English only today | Often many |
| Performance | Depends on your device | Depends on the provider |
| Cost | Free, no account | Usually subscription or per-minute |
“Usually” matters: cloud tools differ. Check the provider's privacy page and offline behavior before you trust it with a private document.
Why local matters
A private note, an unpublished draft or a library book should not need an upload before it can be heard. Local speech keeps the file in the same tab as the text. It also means there is no per-minute voice fee and no account to manage.
Local speech is not automatically better at everything. A cloud service may offer more languages, more voices or faster processing on a weak device. The trade-off is that the file usually goes to a server.
What local does not mean
- It does not mean no network at all. The first visit downloads the app, its speech runtime and the voice.
- It does not mean the host sees nothing. Standard request metadata is still processed.
- It does not mean every device performs the same. Speed depends on the device and browser.
- It does not mean multilingual. The voice is English-only today.
- It does not mean an audiobook file. Generated audio is played as it is made, not saved.
Local, on-device and offline are different claims
Local describes where speech runs. On-device describes what happens to your file. Offline describes what still works without a connection. Read the privacy policy for the full data flow, and what works offline for the setup steps.
Limits
- English only today.
- DRM-free EPUB and text-layer PDFs only.
- No OCR for scanned PDFs.
- No audiobook export in the current UI.
- Performance depends on your device and browser.
Questions
Short answers to the things people ask before opening a file.
Does the voice really run on my device?
Yes. The natural-sounding neural voice runs in your browser. Local files are read and narrated there, and there is no cloud voice service.
Is my book uploaded?
No for local files. The file is read in the browser and is not uploaded. A URL import is a direct request from your browser to that article.
Does local mean offline?
Not automatically. Local is where speech runs. Offline is what still works after the app and voice have been cached. The first visit needs internet.
Do I need an account?
No. There is no account, subscription or per-minute fee.
Does local speech work on a phone?
Yes, in a current browser. Performance varies by device, so a slower phone may take longer to prepare the voice or generate audio.
Is the local voice open source?
No. The application source is not publicly distributed. The local voice is a third-party component with its own license.
Try a local voice with a private file
Open a DRM-free EPUB or a text-based PDF. The file stays in your browser.