Parsing is simply the process of taking apart a document and turning it into structured information that a computer can understand.

When you upload a file—like a PDF, Word document, or even a scanned image—the raw content isn't always in a form that an AI model can easily read. For example, text in a PDF might be stored in fragments, out of order, or even as part of an image.

Parsing is the step that:

• Extracts the text from the document.
• Identifies structure such as paragraphs, headings, tables, and lists.
• Organizes the content so it can be searched, analyzed, or fed into an AI model.

Think of parsing like translating a messy document into a clean, machine-readable format. Without it, the AI might miss key information, read text in the wrong order, or ignore data entirely.

In short: Parsing is how Model HQ makes sure your documents are cleanly understood before any AI tasks—like answering questions, summarizing, or running analysis—are applied.

Document Parsing Options

Model HQ provides multiple parsing modes to ensure accuracy and speed across different types of documents.

• Native Parser (Default)
  The native parser is the fastest option and works extremely well for the majority of text-based documents. It is optimized for performance and should be used as the primary method whenever possible.
• OCR Parsing
  Some documents may be image-based (such as scanned PDFs or files with embedded text as images). In these cases, Optical Character Recognition (OCR) parsing is required to accurately extract text.
• Vision Model Parsing
  For documents that include many images, complex layouts, or require multimodal understanding, you can use a Vision model. This option leverages advanced AI vision capabilities to interpret both text and visual content.

How to Select Parsing Options for RAG use cases for Chat Interfaces

1. In Chat or Bot: Click on the ⚙ icon below the chat box.
2. In the RAG + Generation Config Options, choose between:
   • Native Parser (default, fastest)
   • OCR (for image-based documents)
   • Vision Model (for documents rich in images or requiring visual context)

Conclusion

In this documentation, we explored the RAG source-building functionality in Model HQ.
We covered how to create new and load existing RAG sources, and utilize tools like document upload, semantic search, testing, and configuration settings.

If you have any questions or feedback, please contact us at support@aibloks.com.