AI Lab

Every demo here is a real model call, with its latency and cost shown.

Working demos, not screenshots. Every result below is computed while you watch — in this page, or by this site's own API.

• Tokenizer playground

  See how Persian and English tokenise, and what that costs.

• RAG vs raw

  The same question answered with and without retrieval.

• Semantic vs keyword search

  One query, two engines, side by side over my own corpus.

• Embedding explorer

  Type phrases and watch them find their place in a live 2D projection.