WikiReader is a lightweight web app that turns Wikipedia articles into spoken audio using either
offline browser based speechSynthesis or OpenAI's gpt-4o-mini-tts. It fetches articles via the
official Wikipedia REST API, parses them and reads them aloud in near real-time.
- Choose your playback option.
- Paste a Wikipedia article URL and click Fetch Article.
- The app fetches the main article text via the Wikipedia public API and parses it.
- Uncheck "Read introduction only" if you want the whole article to be read.
- Browser's build-in support is local and free. For OpenAI you need to provide your own OpenAI API key. Note: API key is only used for TTS request and is not stored. Cost of TTS depends on the article length and is billed directly by OpenAI from your account.
- Run
npm installto install dependencies. - Run
npm run devto start the development server.
This project is built with SvelteKit and TypeScript and is under active development. Expect bugs!
Current plans include a text-to-speech pipeline, Wikipedia article fetching and parsing, audio caching, usage tracking, login with basic/full tiers, and production-ready deployment.