I wanted a read-aloud extension that could read from the Kindle Online Reader . The ones I tried used voices I did not enjoy listening to or had a UX that was too cluttered or annoying. A good enough option that I found is CastReader but it requires a subscription and the UI is somewhat buggy. I waned something that runs locally or with a preferred cloud provider.

Now since I have some AI subscriptions, I decided to vibe code my own. In about 2 evenings I have a tool that is fairly usable and serves my purpose. I call it Narro . It reads the page you’re on, highlights each sentence as it’s spoken, and works with most modern AI models. It supports both local and cloud hosted models - the local models run in browser using WebGPU for improved performance.

What it does

• Highlights each sentence as it’s read, and scrolls to follow along.
• Sends several sentences at a time, so the voice keeps a natural rhythm instead of resetting at every full stop.
• Lets you start anywhere: click Start from and pick a sentence, or select text and right-click to start reading from there.
• Remembers where you stopped on each page.
• Skips half-sentences cut off at the edge of a page, and joins a sentence that continues on the next page.
• Can be paused or stopped from any tab. The toolbar icon shows a red ▶︎ while it’s playing.
• Reads Kindle Cloud Reader books using on-device text recognition (uses Tesseract.js )
• It can turn pages inside the Kindle Cloud Reader as it goes.

Voices

• In browser: Kokoro (the default), Supertonic and Piper, Kitten TTS. Models are downloaded on the fly and then cached. Everything works locally.
• Cloud: Supports OpenAI, ElevenLabs, Gemini, or OpenRouter. The user needs to provide the respective api key for the cloud models to work.
• Other: It supports OpenAI-compatible speech server running locally. I have not tested this.

I have open sourced the tool, have a look: GitHub