Drop a PDF onto the Picsart Flow canvas and a Document node is created. It is a new input type, and it hands the file to the rest of your workflow three different ways: as extracted text, as the page you have open, or as every page at once.<\/p>\n
The latest version of Flow is live now, and it adds document input plus two tools for working with large canvases. Together they make it possible to build pipelines directly from PDFs, find any node instantly, and organize hundred-node graphs by meaning rather than position.<\/p>\n
Here is what each one does.<\/p>\n
Until now a workflow started from a prompt or from media. It can now start from a document.<\/p>\n
Add a PDF the same four ways you already add an image:<\/p>\n
Any of those creates a Document node holding the file.<\/p>\n
The Document node is the first node in Flow with three output lanes. Which lane you wire from decides what the rest of the workflow receives.<\/p>\n
Each lane has its own color and shows its name on hover, so wiring stays clear.<\/p>\n
The node is not only a file holder. Inside it you can:<\/p>\n
That last one matters on a canvas you have already built out. The file changes and the wiring does not.<\/p>\n
The Document node is a static input. It uses no credits and it does not run in the pipeline.<\/p>\n
All processing happens in the browser. PDFs up to 50 MB and 100 pages are supported. Corrupted or damaged files show a clear error, and scanned PDFs without text show a hint.<\/p>\n
Finding a node on a large workflow no longer means panning and squinting. Press Cmd\/Ctrl+F anywhere on the canvas to open the search bar.<\/p>\n
Sticky notes are full search results, not a secondary category, and they are highlighted the same way as nodes.<\/p>\n
Search is read-only, so it is also available to collaborators with view-only access. On a 100-node workflow it finds and centers a node in under a second.<\/p>\n