๐Ÿค– RAG-Powered Document Intelligence

Chat with your Documents

Upload any PDF or text file. Ask questions in natural language. Get answers grounded in your document's actual content โ€” not AI hallucinations.

๐Ÿ“„
PDF Upload
๐Ÿง 
Smart Chunking
๐Ÿ”
Vector Search
๐Ÿ’ฌ
AI Answers

Powered by a Full RAG Pipeline

๐Ÿ“ฅ

Document Ingestion

Upload PDF or plain text files up to 20 MB. The system extracts text from every page automatically.

โœ‚๏ธ

Recursive Chunking

Text is split using RecursiveCharacterTextSplitter โ€” preserving paragraphs & sentence boundaries with 200-char overlap.

๐Ÿงฌ

Google Embeddings

Each chunk is embedded with text-embedding-004 โ€” Google's high-quality embedding model.

๐Ÿ—„๏ธ

Qdrant Vector DB

Embeddings are indexed in Qdrant for blazing-fast similarity search with cosine distance.

๐Ÿ”Ž

Semantic Retrieval

Your question is embedded and matched against the top-5 most relevant chunks from the document.

๐Ÿค–

Grounded Generation

Gemini 2.0 Flash generates answers strictly from the retrieved context โ€” not from its general knowledge.

How It Works

1

Upload Your Document

Drag & drop or click to select a PDF or TXT file. The system accepts files up to 20 MB.

2

Automatic Processing

The document is chunked, embedded, and indexed into a vector database in seconds.

3

Ask Questions

Type any question about your document in natural language.

4

Get Grounded Answers

The AI retrieves the most relevant sections and generates an accurate answer with page references.