Yes. The free tier reads any web page, PDF, or selected text using a neural AI voice that runs on your own computer. Unlimited reading, no daily cap, no account, no trial period. An optional Pro upgrade ($29.99 one-time, not a subscription) lets you add your own API key to use premium cloud voices from providers like ElevenLabs, OpenAI, and Cartesia.
AI Voice Reader works with Google Chrome and Chromium-based browsers (Edge, Brave, Arc, Opera, Vivaldi). It uses Chrome’s WebGPU and offscreen-document APIs, so it does not work in Firefox or Safari.
Yes, on the free tier. Once the neural voice model is downloaded the first time you install it, all reading happens entirely on your computer with no internet connection needed. The Pro tier uses your chosen cloud provider, which requires the internet.
The free tier uses Kokoro, a small open-source neural voice model. It’s about 90 MB and downloads automatically once, the first time you install the extension. After that, every page you have read aloud is synthesized locally on your machine. The download is one-time; the voice quality is much better than the robotic voices built into your browser.
Pro lets you add your own API key from a premium voice provider (ElevenLabs, OpenAI, Cartesia, Deepgram, Azure, Google Cloud, or AWS Polly) and use those voices directly inside the extension. You pay the provider for what you actually use, on their pricing. We never see your usage and never charge you per word. Pro is a one-time $29.99 payment, never a subscription.
Cloud voices cost money to generate. Subscription read-aloud apps mark up that cost and gate usage behind monthly caps to manage their margin. Bring-your-own-key removes the middleman: you pay the provider their actual rate, get unlimited use within their pricing, and we have nothing to refund or cancel.
Yes. AI Voice Reader uses whichever voices your provider account makes available. If you set up a custom or cloned voice on your provider (for example, on the ElevenLabs website), it will appear automatically in the voice picker the next time you reload the extension. We do not handle voice cloning ourselves; that relationship is between you and your provider.
Articles, blog posts, documentation, emails in Gmail, Google Docs, PDFs opened in the browser, and any text you highlight on a page. Click the toolbar icon to read the current page, or right-click selected text to read just the selection.
It works on most reading-focused sites. Kindle Cloud Reader, Pocket-style readers, and most article sites are supported. A small number of sites with heavy iframe nesting or unusual DOM structures may not work; report any specific site that fails and we will look at it.
On the free tier, no. All reading happens locally and no page content leaves your computer. On the Pro tier, only the text you choose to read is sent directly from your browser to whichever cloud provider you configured, using your own API key. We never see the text and there is no analytics middleman.
Install it from the Chrome Web Store. Click “Add to Chrome” and the extension is ready immediately, using your browser’s built-in voices. The higher-quality neural voice model downloads in the background and takes over the next sentence after it finishes.
The 90 MB neural model download is one-time and runs in the background; while it downloads, the extension uses your browser’s built-in voices so reading works immediately. If you would prefer to skip the neural voice entirely, you can delete the downloaded model from Settings and the extension will stay on the browser’s built-in voices forever.
If you have Enhanced Safe Browsing enabled in Chrome, you may see a “Proceed with caution” warning when installing. This is normal for newer extensions on the Chrome Web Store. Google’s system requires developers to build a track record of compliance over several months before granting “trusted” status. AI Voice Reader requests only the minimum permissions needed and synthesizes voices entirely on your device on the free tier. You can safely click “Continue to install.” For more details, see our Privacy Policy.