README.md

June 6, 2026 Β· View on GitHub

πŸ‘οΈ SightlineAI

AI-Powered Accessibility Guidance Platform for the Visually Impaired

Empowering independent navigation through multimodal artificial intelligence.

"🌐 Live Demo" (https://rudra496.github.io/sightlineai) β€’ "πŸ“– Documentation" (https://github.com/rudra496/sightlineai) β€’ "πŸ› Report Issue" (https://github.com/rudra496/sightlineai/issues) β€’ "πŸ’‘ Feature Request" (https://github.com/rudra496/sightlineai/issues)

"License" (https://img.shields.io/badge/License-MIT-green.svg) "Python" (https://img.shields.io/badge/Python-3.11+-blue.svg) "FastAPI" (https://img.shields.io/badge/FastAPI-Modern_API-009688.svg) "Open Source" (https://img.shields.io/badge/Open%20Source-Yes-success.svg) "AI" (https://img.shields.io/badge/AI-Gemini%20%7C%20Qwen-purple.svg)

---

🌟 Overview

SightlineAI is an open-source accessibility platform designed to help blind and visually impaired individuals better understand, navigate, and interact with their surroundings.

By combining computer vision, OCR, conversational AI, and navigation intelligence, SightlineAI transforms ordinary devices into intelligent accessibility companions capable of delivering real-time environmental awareness and guidance.


πŸš€ Key Features

  • πŸ‘οΈ Real-time scene understanding
  • πŸ“„ OCR and document reading assistance
  • ⚠️ Obstacle and hazard awareness
  • πŸ—ΊοΈ Context-aware navigation support
  • πŸ’¬ Conversational AI assistance
  • 🌐 Multi-language accessibility support
  • πŸ”„ Offline fallback capabilities
  • ⚑ Real-time WebSocket communication

🧠 AI Architecture

SightlineAI follows a multi-model architecture that supports multiple AI providers.

Supported Models

Provider| Capability Google Gemini| Multimodal reasoning, scene understanding, accessibility guidance Qwen| Vision analysis, conversational intelligence, contextual assistance

This provider-agnostic design improves flexibility, reliability, and future scalability.


πŸ—οΈ System Workflow

Camera / Voice Input ↓ Visual & Context Processing ↓ Gemini / Qwen AI Analysis ↓ Risk & Accessibility Assessment ↓ Voice and Text Guidance


🌍 Impact

SightlineAI addresses accessibility challenges faced by more than 285 million visually impaired people worldwide, particularly in regions where advanced assistive technologies remain financially inaccessible.

United Nations SDGs

  • πŸ’š SDG 3 – Good Health & Well-Being
  • ❀️ SDG 10 – Reduced Inequalities
  • πŸ’› SDG 11 – Sustainable Cities & Communities

πŸ› οΈ Technology Stack

Category| Technologies Backend| Python, FastAPI Frontend| React, HTML, CSS, JavaScript AI Models| Google Gemini, Qwen Vision| OCR, Computer Vision Infrastructure| Docker, Redis Database| SQLite, PostgreSQL Navigation| OpenStreetMap


βš™οΈ Quick Start

git clone https://github.com/rudra496/sightlineai.git

cd sightlineai

cp .env.example .env

pip install -r requirements.txt

python -m app.main


🀝 Contributing

Contributions are welcome.

Areas of contribution include:

  • Accessibility improvements
  • AI model enhancements
  • Documentation
  • Localization
  • Bug fixes
  • UI/UX improvements

πŸ“„ License

Released under the MIT License.


πŸ‘¨β€πŸ’» Maintainer

Rudra Sarker

Industrial & Production Engineering Shahjalal University of Science and Technology (SUST)

🌐 Portfolio: https://rudra496.github.io/site

πŸ™ GitHub: https://github.com/rudra496

πŸ’Ό LinkedIn: https://www.linkedin.com/in/rudrasarker

πŸ“§ Email: rudrasarker130@gmail.com


Built for accessibility. Built for independence. Built for impact.