Offline AI education platform for low-resource learners using Ollama, WebLLM, React & Edge AI.
# π Offline AI Chatbot β Low-Resource AI for Everyone
**A fully offline AI chatbot platform built to close the Digital Divide β running powerful language models locally on any device, with zero internet dependency.**
π Quick Start Β· π How It Works Β· π οΈ Tech Stack Β· πΊοΈ Roadmap
---
## π― Project Overview
**Offline AI Chatbot** is a full-stack AI chat application that enables students, educators, and users in low-resource environments to interact with large language models **entirely offline** β no cloud subscription, no high-speed internet, no data privacy concerns.
It supports **two AI inference modes**:
- **Ollama** β runs local LLMs natively on device through a backend Express server
- **WebLLM** β executes AI models directly in the browser using WebGPU, with no server required
Built as part of the **"Digital Divide 2.0: Low-Resource AI"** academic initiative by **Team Dev Vengers**, this platform addresses a critical equity gap: most AI tools today require high-end hardware and reliable internet β leaving behind rural students, low-income learners, and schools in under-served communities.
---
## β Problem Statement
The modern AI revolution is creating a new kind of inequality β the **AI Intelligence Gap**:
- π Most AI tools require **high-speed internet** and **cloud-based processing**
- π» Leading AI platforms demand **high-end hardware** that most students cannot afford
- π« Schools in rural or low-income areas are systematically **excluded from AI-powered education**
- π Cloud AI tools raise **privacy concerns** about student data sent to third-party servers
While AI-enabled students surge ahead, millions of learners are left behind β not because of lack of intelligence, but because of lack of access.
---
## β
Proposed Solution
An open-source, **fully offline AI chatbot platform** that:
- Runs **state-of-the-art LLMs locally** on the user's own device
- Works in **zero-internet environments** β no Wi-Fi, no data plan needed
- Supports β¦