ListenLM

Formats AI chat responses for screen readers with ARIA landmarks, code block labels, and keyboard navigation.

ListenLM adds screen reader accessibility to Claude, ChatGPT, Gemini, and Copilot. It labels code blocks by language, marks headings, tables, lists, and quotes with proper ARIA roles, announces when responses finish generating, and adds keyboard shortcuts to jump between code blocks and headings within a response, all without changing how the page looks. If a chatbot outputs: ## Setup - Install the package - Configure your API key ```javascript npm install foo const key = process.env.API_KEY Without ListenLM: a screen reader hitting a typical AI response just reads the visible text flatly, with no structure cues: "Setup Install the package Configure your API key npm install foo const key = process dot env dot API underscore KEY" With ListenLM: the same response gets structural landmarks announced: "Heading: Setup. Bulleted list, 2 items. Bullet: Install the package. Bullet: Configure your API key. JavaScript code block. npm install foo const key = process dot env dot API underscore KEY. End JavaScript."