Portfolio

Talk to Docs is a multilingual, AI-powered application that lets users chat with their documents naturally — through both text and voice input. It was built using LangChain for orchestration, ChromaDB as the vector database, and LLaMA 3 as the open-source LLM backbone.

The system allows users to upload PDFs, Word files, or text documents and interact with them conversationally. Queries can be made in different languages or through voice, making the tool accessible for global users. Behind the scenes, Talk to Docs processes documents, embeds them into a vector space with ChromaDB, and uses retrieval-augmented generation (RAG) to deliver context-aware, real-time answers.

By combining multilingual support, voice input, and open-source AI models, Talk to Docs provides an efficient and cost-effective solution for students, researchers, and enterprises. Early adopters reported up to a 70% reduction in document review time, highlighting its value as a productivity enhancer.