On-Device Text-to-Speech Chrome Extension: Read Pages Without Uploading Text

On-Device Text-to-Speech Chrome Extension: Read Pages Without Uploading Text

An on-device text-to-speech Chrome extension reads a web page by turning text into audio in the browser, so that page text does not go to a cloud TTS API to generate speech. That is the product type to install if your requirement is “do not upload the article.” Flow TTS is built that way for supported models. It is not zero-cloud: accounts, usage, licenses, and model downloads still use the network. If that boundary matches what you want, install it and try Auto mode on one clean article.

What does on-device mean in a Chrome extension?

On-device, or local, synthesis means the speech model runs on this computer. The extension collects readable text from the page, the model writes audio here, and the in-page player plays it. A cloud TTS extension does the middle step on a server: it sends the text out and gets audio back.

That is a data-flow choice, not a quality guarantee. A local model can sound natural or thin. A cloud API can sound excellent and still upload every paragraph. For the architecture tradeoffs — latency, hardware, voice catalogs — read local vs cloud text-to-speech. For how a local neural model actually runs in the extension — offscreen worker, ONNX, WebGPU or WASM — read how Flow TTS generates speech locally. This page is only about what to install when “page text stays off a TTS API” is the requirement.

On-device is also not the same as offline. A Chrome extension can synthesize locally and still need the network for sign-in, usage, or downloading a model the first time.

How do you tell if an extension uploads page text?

Marketing pages blur this. Check the product, not the adjective “private.”

1. Where is the audio generated? If speech needs a vendor TTS API on every play, page text is leaving the device for synthesis. If the extension runs an installed or bundled model in the browser, synthesis can stay local.

2. Does the privacy policy match the settings? A policy that lists a third-party speech vendor is a cloud path. A policy that separates “speech stays on-device” from “accounts and downloads use the network” is the honest local shape. Read the Flow TTS privacy policy against that test.

3. Is a model download different from a per-play upload? Downloading a voice engine once is not the same as sending each article to a server. Ask whether later playback still needs that API.

4. What is stored, and where? Local synthesis does not mean nothing is stored. Accounts, daily usage, and license state often live on a backend even when audio does not.

If you cannot answer those four, do not treat the extension as on-device. For a longer shopping list that includes player controls and usage caps, use what to look for in a text-to-speech Chrome extension.

What should a good on-device reader still do?

Keeping text off a TTS API is necessary. It is not sufficient.

• Read the article, not the chrome. The tool should collect main readable content. “Full page” usually means menus and footers in the queue.
• Read a selection. A right-click or highlight path matters when you only want one section.
• Stay on the page. An in-page player with pause, skip, and speed is the difference between a demo and a reader.
• Let you add engines. Local catalogs are finite. You should be able to install another model without switching to a cloud API.

If you are still deciding how to listen at all, start with how to listen to web pages in Chrome.

Where Flow TTS fits

Flow TTS is an on-device text-to-speech Chrome extension for reading web pages. Speech synthesis runs in the browser for supported installed or bundled models, so page text used for audio does not need a cloud TTS service.

What that looks like in the current product:

• Install from the Chrome Web Store listing, then pin the toolbar icon.
• Bundled Supertonic v1 is ready without a Models-tab download. Other families — Kitten, Kokoro, Piper, Supertonic v2, and TinyTTS — install from Settings → Models. The registry has 27 model IDs across those five families. Finish an install before you play that model.
• Auto mode reads readable main content on the current page. It is not a promise that the full layout will be spoken.
• Selected text uses the context-menu read-selection action.
• Flow Templates save Read and Skip targets for sites you revisit.

Limits that belong on this page, not after you install:

• Play requires an account signed in with an email one-time code. You can continue the onboarding tour without signing in. Playback will not start until a token is present.
• Free usage is 10,000 characters per day. Premium unlocks unlimited characters and Glass Mode when entitlement is active.
• Accounts, license checks, usage, and model downloads use the network. That is why Flow TTS is not fully offline and not zero-cloud.

First-session steps are in Getting Started with Flow TTS.

When should you pick a cloud tool instead?

Pick cloud TTS if you need a language or a specific hosted voice that a local catalog cannot cover, or if you want the laptop to do none of the synthesis work. Those are real reasons. They are also the cases where page text will leave the device for speech. Do not install a local extension and then expect a cloud-sized voice list on day one.

If on-device is the requirement, install a local reader, try one clean article, and confirm the four checks above against its settings and privacy policy.

FAQ

Does an on-device text-to-speech Chrome extension upload page text?

Not for synthesis, if it is actually on-device. Page text is turned into audio in the browser. Other features — accounts, usage, licenses, model downloads — may still use the network.

Is Flow TTS fully offline?

No. Speech can run locally for supported installed or bundled models. Sign-in, usage checks, license checks, and downloading models can require a connection.

Do I need an account to use Flow TTS?

Yes. Play needs a signed-in account. Free users get core reading after email one-time-code setup, subject to 10,000 characters per day.

Is this the same as the local vs cloud guide?

No. That post explains the two architectures. This post is for choosing and installing an on-device Chrome extension.

How do I try Flow TTS?

Install it from the Chrome Web Store, complete account setup, start with bundled Supertonic, and use Auto mode on a clean article.