Import a real document
Start with a PDF, DOCX, or DOC file. HyprDuck keeps the primary flow file-first, without screen recording or accessibility permission friction.
HyprDuck imports PDFs and Word files, renders every page into images, runs your chosen AI provider, and saves linked markdown that stays inspectable on your machine.
HyprDuck keeps the landing page and product surface centered on the same path: file in, page evidence, provider run, linked markdown out.
Start with a PDF, DOCX, or DOC file. HyprDuck keeps the primary flow file-first, without screen recording or accessibility permission friction.
Each page becomes an inspectable image snapshot before AI processing, so the markdown can keep a visible link back to the source page.
Use local mode or a hosted provider depending on your privacy, speed, and cost needs. Provider settings stay shared across parsing flows.
HyprDuck writes markdown and linked image assets under Application Support, ready for agents, wikis, notes, and local knowledge workflows.
HyprDuck treats AI settings as shared parsing infrastructure. Local operation stays visible, while hosted providers remain available when they fit the job.
Run private document parsing against local models without an API key.
Use one provider setting to reach multiple hosted model families.
Every result should be movable, readable, and grounded in source artifacts on disk.
Document import and page rendering happen in the desktop app before any provider call.
OpenRouter hosted models and local Ollama models share the same parsing workflow.
Generated markdown references the page images on disk instead of hiding the source.