Chalkboard — OCR & Vision Intelligence Platform
Chalkboard reimagines OCR as an interactive, extensible, and accessible vision intelligence system rather than a static text extraction tool. Designed from the ground up for blind and low-vision users.
Key Features
- Dual-Architecture OCR: Offline engines + state-of-the-art AI vision models working together.
- AI Refine: Transform raw text into structured, accessible HTML (headings, tables, lists).
- Smart Compare: Monitor screens and announce only visual changes.
- Queryable OCR Workspace: Ask natural language questions about any visible element.
- Extensive AI Ecosystem: Support for Gemini, OpenAI, Claude, DeepSeek, Grok, and more.
- Fully Accessible GUI: Keyboard-first, screen-reader optimised, WCAG 2.1 AA/AAA targeting.
Keyboard Shortcuts
| Action | Shortcut |
|---|---|
| OCR Capture to Workspace | Alt + 1 (or Alt + Ctrl + Z) |
| Set Smart Compare Baseline | Alt + 2 (or Alt + Ctrl + X) |
| Compare with Baseline | Alt + 3 (or Alt + Ctrl + C) |
| AI Refine (inside workspace) | Ctrl + Shift + R |
| Query (inside workspace) | Ctrl + Enter |
| Settings | Ctrl + , |
AI Provider Support
Google Gemini, OpenAI GPT-4o, Anthropic Claude, DeepSeek, xAI Grok, Perplexity, Qwen
Compatibility
Python 3.9+ | Windows 10 and later | Screen-reader optimised (NVDA/JAWS)
Developed by: Susant Swain, Himanshu Ranjan, Krishna Dubey — Team 4 for Vision-Aid "Hack for Inclusion" 2026.