Inspiration
Many citizens struggle to discover and understand government schemes due to language barriers and inaccessible formats.
Purpose
This assistant enables users to speak in their native language and receive spoken answers about government schemes, fetched live from official sources.
How It Was Built
Built with Next.js, Node.js and deployed on Tencent EdgeOne for low-latency performance. It uses LlamaIndex + Pinecone for semantic retrieval, LangChain and Hugging Face Transformers for intelligent Q&A, and integrates speech-to-text, text-to-speech, and OCR for full voice interactivity. Users can also upload PDFs or scanned documents to chat with them directly, making scheme discovery accessible even from offline sources.
Target Users/Scenario
Designed for global impact—citizens worldwide, especially in rural or multilingual regions, who need accessible, conversational access to government benefits and schemes.