2025

Document RAG Service

A Retrieval-Augmented Generation API for querying your own documents.

• NestJS
• TypeScript
• ChromaDB
• LangChain
• OpenAI / OpenRouter
• Docker

• Built a RAG API that processes uploaded PDF, DOCX, and TXT files, chunks and embeds them into ChromaDB, and answers natural-language questions using an LLM.
• Leveraged LangChain's RecursiveCharacterTextSplitter for intelligent document chunking and OpenRouter for flexible LLM provider access (Meta Llama 3).
• Fully documented via Swagger; containerised with Docker Compose for local vector database management.