👁️ Visual UI to debug document chunking and extraction

[CharnaParkey]

We're considering building a visual document extraction viewer to help you fine-tune your ingestion pipeline.

What it would show:

• Bounding boxes around extracted chunks
• Mapping between visual layout and ingested content
• Chunk boundaries overlaid on the original document
• Metadata extraction visualization

Use cases:

• Debug why certain content isn't being retrieved correctly
• Understand how your chunking strategy affects document structure
• Fine-tune chunk size and overlap settings
• Verify OCR and extraction quality

Example workflow:

1. Upload a document
2. See exactly how OpenRAG is chunking it
3. Adjust settings and see results in real-time
4. Export optimized configuration

Questions for the community:

• What document types do you struggle with most?
• Would you want this in the UI or as a separate tool?
• What other information would you want visualized?

Vote with 👍 if you'd use this feature!

Tell us about your document extraction pain points below.