Uploaded July 2026 | Updated November 2026, 14 hours ago
A scanned PDF and a native electronic PDF present different extraction problems for LLMs. Most tools treat them the same.
At DevCon 6, this AIP Workflow Optimization demo launches the new Text Extract V2 board, used inside a logic function to take a PDF media reference as input and return either raw text or a structured representation of the document. Configurable properties let you tune extraction for electronic PDFs versus scanned PDFs, and control the output format: raw text and markdown, both supported since Text Extract V1, plus a new HTML output for highly structured responses.
A scanned PDF and a native electronic PDF present different extraction problems for LLMs. Most tools treat them the same.
At DevCon 6, this AIP Workflow Optimization demo launches the new Text Extract V2 board, used inside a logic function to take a PDF media reference as input and return either raw text or a structured representation of the document. Configurable properties let you tune extraction for electronic PDFs versus scanned PDFs, and control the output format: raw text and markdown, both supported since Text Extract V1, plus a new HTML output for highly structured responses.
