Skip to content
Pitaj.io

Knowledge Base Overview

The Knowledge Base is the core foundation of your AI agent. Unlike generic ChatGPT prompts that have limited context windows or risk hallucinating, Pitaj.io uses an enterprise semantic RAG (Retrieval-Augmented Generation) pipeline:

  1. Your source materials are parsed and segmented into semantic chunks.
  2. Each chunk is vectorized and stored in our secure, isolated vector database.
  3. When a user asks a question, the system retrieves the most relevant chunks in milliseconds and provides them to the model to generate a factual, citation-backed response.

Supported Knowledge Source Types

Websites (Web Crawler)

Enter your domain or sitemap URL. Our automated crawler indexes all subpages and periodically syncs updates. Learn more →

Documents & Files

Upload PDF product catalogs, Word manuals, TXT notes, or pricing sheets. Text is automatically parsed and indexed. Learn more →

Text Snippets

Quickly paste custom policies, holiday working hours, return guidelines, or internal instructions. Learn more →

Questions & Answers (Q&A)

Define explicit question-and-answer pairs for high-stakes topics that require exact, non-negotiable responses. Learn more →