_Th---
title: SubText ---
title: SubText — Real-Time Code-to-Speech Accessibility for Blind Developers
published: true
tags: mlhacks, devchallenge, accessibility, webdev
This is a submission for the MLH x DEV Writing Challenge.
🎙️ What I Built
Screen readers are incredible, but when it comes to programming code, they are brutally literal. Hearing a screen reader chant "open parenthesis const curly brace arrow semicolon" at 400 words per minute is a massive mental drain for visually impaired developers or developers with cognitive processing differences.
SubText is a browser extension and open-source accessibility overlay that translates raw, syntax-heavy code on websites like GitHub, StackOverflow, or LeetCode into plain-English conversational logic—and reads it out loud using natural Speech Synthesis.
Instead of reading raw punctuation, SubText explains what the code does:
Raw Code:
if (!user || user.status !== 'ACTIVE') return null;
SubText Reader: "If the user doesn't exist or isn't active, exit early and return null."
🎥 Demo
!embed
(Replace with your YouTube / Loom walkthrough URL, or embed a GIF/Image here)
- 🌐 Live Extension / Web App: https://subtext-demo.vercel.app (Replace with your link)
- 💻 GitHub Repository: https://github.com/your-username/subtext
💡 Inspiration
During a late-night hacking session, I watched a visually impaired software engineer try to debug a complex nested loop using a traditional screen reader. The sheer amount of verbatim punctuation noise made following the control flow exhausting.
Code syntax was designed for human eyes, but voice control for code is usually an afterthought. I wanted to build a tool that treats code accessibility not as literal OCR reading, but as semantic translation.
🛠️ How It Works & Tech Stack
SubText acts as an intelligent intermediary between DOM code elements and browser audio synthesis.
— Real-Time Code-to-Speech Accessibility for Blind Developers
published: true
tags: mlhacks, devchallenge, accessibility, webdev
This is a submission for the MLH x DEV Writing Challenge.
🎙️ What I Built
Screen readers are incredible, but when it comes to programming code, they are brutally literal. Hearing a screen reader chant "open parenthesis const curly brace arrow semicolon" at 400 words per minute is a massive mental drain for visually impaired developers or developers with cognitive processing differences.
SubText is a browser extension and open-source accessibility overlay that translates raw, syntax-heavy code on websites like GitHub, StackOverflow, or LeetCode into plain-English conversational logic—and reads it out loud using natural Speech Synthesis.
Instead of reading raw punctuation, SubText explains what the code does:
Raw Code:
if (!user || user.status !== 'ACTIVE') return null;
SubText Reader: "If the user doesn't exist or isn't active, exit early and return null."
🎥 Demo
!embed
(Replace with your YouTube / Loom walkthrough URL, or embed a GIF/Image here)
- 🌐 Live Extension / Web App: https://subtext-demo.vercel.app (Replace with your link)
- 💻 GitHub Repository: https://github.com/your-username/subtext
💡 Inspiration
During a late-night hacking session, I watched a visually impaired software engineer try to debug a complex nested loop using a traditional screen reader. The sheer amount of verbatim punctuation noise made following the control flow exhausting.
Code syntax was designed for human eyes, but voice control for code is usually an afterthought. I wanted to build a tool that treats code accessibility not as literal OCR reading, but as semantic translation.
🛠️ How It Works & Tech Stack
SubText acts as an intelligent intermediary between DOM code elements and browser audio synthesis.
is is a submission for the MLH x DEV Writing Challenge_
Top comments (0)