A progressive, script-based journey through Retrieval-Augmented Generation, from basic RAG to agentic, self-reflecting systems. Free-tier cloud stack, zero local model downloads.
Agentic RAG with LangGraph, guardrails, FlashRank reranking, and page-accurate citations. Runs locally on Groq + embedded Qdrant.
Production-grade document intelligence and RAG platform. Users upload PDFs, DOCX, PPTX, XLSX, CSV, etc., then ask questions and receive cited answers. Uses hybrid vector + keyword retrieval, reranking, agentic retrieval retries, conversation memory, and LLM evaluation.
Supplementary benchmarks for Making Legacy Knowledge Searchable with RAG
AI-powered hybrid search engine for arXiv papers. Natural language queries → arXiv API → FAISS + BM25 → RRF fusion → FlashRank reranking → LLM synthesis.
Local ONNX cross-encoder reranking for JavaScript/TypeScript. Zero API costs, framework-agnostic, with a Vercel-style API.
A low-latency full-stack RAG chatbot built with FastAPI, React, PostgreSQL, and pgvector, featuring PDF/DOCX ingestion, hybrid retrieval, grounded chat, streaming responses, and optional FlashRank reranking.
AI-Powered NLP Dataset Generator & Semantic Search Platform | FastAPI, Next.js, Redis Vector DB, FlashRank, Docker
To associate your repository with the flashrank topic, visit your repo's landing page and select "manage topics."