Developed a local demo API service for the Shubhamsaboo/LightRAG repository, enabling embeddings and model completion workflows through FastAPI endpoints for querying, text insertion, and file ingestion. The implementation integrated the LightRAG framework with a local Ollama instance, leveraging Python for backend logic and API development. This prototype established a reproducible environment for rapid experimentation with content ingestion and retrieval, supporting efficient validation of embeddings and model-completion features. The work included code formatting improvements to enhance maintainability. By focusing on LLM integration, RAG techniques, and robust API design, the project accelerated testing and demonstration of advanced retrieval-augmented generation capabilities.