qianfanocr-document-intelligence

To extract text and layout structure from a PDF, use the document parsing with layout mode. It processes visual document pages, preserves structural elements, and returns structured outputs for downstream agent reasoning and automation.

Yes, you can extract key fields from invoice images using the key information extraction mode. It interprets visual document content to identify and return structured data for downstream automation and reasoning.

The best way to perform OCR and chart understanding on images is using a deterministic processing pipeline supporting general OCR and chart understanding modes. It extracts, interprets, and locates visual content to return structured results.

Yes, this document intelligence tool processes both local files and URLs. It accepts images, image URLs, PDFs, and PDF URLs as valid inputs to extract content and enable document visual question answering.

To answer questions about content within a visual document, use the doc VQA mode. It interprets visual inputs, locates relevant content, and returns structured results to upstream agents for reasoning.