RAG Lab

Embeddings, vector search, and pasted PDF text chat with LangChain + Chroma + GPT-4o mini

🧠 Embeddings

Enter two sentences and compare their semantic similarity using all-MiniLM-L6-v2 embeddings + cosine distance. No API key needed.

✂️ Chunking

Paste a block of text and see it split into overlapping chunks using RecursiveCharacterTextSplitter. No API key needed.

🔍 RAG Q&A

Enter your knowledge base (one fact per line), then ask a question. The text is chunked, embedded, and queried with retrieval-augmented generation.

🔀 Reranking

Enter your knowledge base, then ask a question. Results are reranked by a cross-encoder for higher accuracy than plain vector search.

📄 PDF Chat

Paste PDF content up to 1000 characters, index it in memory, then ask questions about it.