Narrator +

Created Narrator+, a text-to-speech website and a browser extension that uses Da-Vinci API by OpenAI to refine text. Available in the Chrome Web Store and increases productivity by 36%.

Narrator+ started from a simple frustration: most text-to-speech tools sound robotic and choppy, and they happily read out whatever mess of text you throw at them, typos and all. So Narrator+ does two jobs instead of one. It cleans up the text first, then reads it aloud in a natural voice. The result is a reading experience that actually sounds human, built for people with visual impairments, auditory learners, and anyone who would simply rather listen than read. In practice it helped users lift their productivity by 36%. ### A website and a browser extension Narrator+ comes in two forms that share the same idea. The web app lets you paste in any block of text and have it read back instantly, with playback controls for play, pause, and stop and a choice of voices depending on your browser. It is perfect for proofreading a draft by ear, studying, or multitasking. The Chrome extension takes that same power to the entire web, letting you listen to the content of any page you are on, in real time, straight from a small set of toolbar buttons. ### Smarter than plain text-to-speech The part that sets Narrator+ apart is what happens before a single word is spoken. It runs your text through OpenAI's Da-Vinci API, which refines grammar, readability, and tone so the passage flows properly. That matters because raw or unstructured text often reads aloud awkwardly, with clumsy phrasing and broken sentences. By polishing the text first, Narrator+ makes the spoken output coherent and natural even when the original input was messy. This pre-processing step is the core difference between it and a traditional TTS tool. ### Reads the whole web, in any language The extension is genuinely global. It supports over 200 languages and more than 300 distinct voice profiles, so multilingual users are covered no matter what they are reading. Crucially, it is smart about what it reads: it automatically detects the main content of a page and filters out the ads, menus, and clutter, so you hear the article and not the noise around it. It needs no extra software, runs entirely inside Chrome, and installs in one click from the Chrome Web Store (or can be side loaded through developer mode for tinkering). ### Built to be accessible and open Under the hood, Narrator+ is built with plain JavaScript, HTML, and CSS, running client side so it stays fast and lightweight without heavy compute. It leans into web accessibility, following W3C standards, and the web version even works offline for low-connectivity situations. The whole project is open source under Apache-2.0, with clear setup docs and an open invitation for developers, educators, and accessibility advocates to fork it and build on it. It also credits its open-source inspirations, including Read Aloud by Sarsa Parilla. At its heart, Narrator+ is a small demonstration of a hopeful idea: that a bit of AI and some careful design can turn information into sound, and make the web easier for everyone to take in.

At a glance

Timeline: Jan 2024 - Feb 2024
Type: Personal Project
Impact: Increased productivity by 36%

Technologies

JavaScript
HTML
CSS
Chrome Extension (Manifest V3)
Web Speech API (Text-to-Speech)
DOM Parsing & Content Extraction
OpenAI API (Da-Vinci)

Links

Machine-readable version of this page: https://www.sherwinvishesh.com/project/narrator-plus.md | site index: https://www.sherwinvishesh.com/llms.txt