PDF Export by PDFCrowd
DocumentsConvert documents between formats with the PDFCrowd HTTP API, rendering web pages and HTML into PDFs or images and extracting text from existing files. Agents generate and read documents in the same connector.
What This Integration Enables
Almost every document connector runs one way. Markup goes in, a document comes out, and the job is done. PDF Export by PDFCrowd is built the other way around as well, and that is the whole reason to choose it. It renders web pages and HTML strings into PDFs or images, and it also reads existing documents: extracting the text layer with the original reading layout intact, converting a PDF into a self-contained HTML file with fonts and graphics embedded, and rasterizing pages into images for previews or downstream vision steps. Contracts, invoices and statements do not travel in one direction, and a connector that only generates leaves you buying a second one the first time something comes back.
The rendering side carries the controls that matter for legal collateral: page size, orientation and margins, HTML headers and footers with automatic page numbering, viewport and scaling control for responsive layouts, custom CSS injection, and wait-for-element handling so a page that assembles itself in JavaScript is captured complete rather than half drawn. The reading side is equally specific. Extraction supports page ranges, layout-free output, paragraph detection, de-hyphenation and empty-line removal, and it can open password-protected files. It reads embedded text only and does not perform OCR, which is a limitation stated plainly by the vendor and one this connector's design should respect rather than paper over. Authentication is a PDFCrowd username and API key, and every conversion is stored in FlowRunner file storage with its URL, filename, format and size returned to the flow.
Without FlowRunner
With FlowRunner
Use Case Scenarios
The Contract Round Trip
The agent renders the agreement from the approved HTML template with Convert HTML to PDF, complete with a footer carrying the matter reference and page numbering, and routes it for signature through DocuSign. Weeks later the countersigned file returns. Now the same connector reads it: Extract Text from PDF pulls the executed language, Convert PDF to Image builds page thumbnails for the matter record, and the agent compares the returned terms against the version it generated. If a clause moved between send and signature, the diff is on the matter record before anyone files the document, not after the first invoice under it is disputed.
Reading Inbound Invoices Into Structured Data
Supplier invoices arrive as PDFs from dozens of senders in dozens of layouts. Extract Text from PDF returns the text with the reading layout preserved, which is what makes vendor names, line items and totals recoverable rather than a scrambled stream. Paragraph detection and empty-line removal clean the output before it reaches the parsing step. The structured result posts against the purchase order in NetSuite, and the extracted text is stored as a .txt file alongside the PDF so the same document never has to be read twice.
Republishing PDF Collateral for the Web
A stack of product datasheets exists only as PDFs, which means they are invisible to site search and awkward on a phone. Convert PDF to HTML produces a single self-contained HTML file per datasheet with images, stylesheets and fonts embedded and the original layout reproduced, and Convert PDF to Image builds a cover thumbnail for the listing card. Convert Image Format then normalizes those thumbnails to web-optimized WebP at a consistent size. The collateral becomes indexable content instead of a download nobody opens on mobile.
Human-in-Loop Highlight
Extract Text from PDF reads the embedded text layer and does not perform OCR on scanned images. That single documented behavior is the sharpest failure mode on this connector, because a scanned countersigned contract returns an empty or near-empty string, and to every downstream step "no unusual termination clause found" and "could not read this document at all" look identical. A clean result and a total failure produce the same green check. So when extraction returns a character count implausibly low for the page count, the agent stops rather than reporting a pass, and messages the contracts manager: "Extraction on [document] returned [N] characters across [M] pages. This is consistent with a scanned document that carries no text layer, so the clause review did not actually run. Page thumbnails are attached. Route this for manual review, or send it through an OCR step first?" That is human-in-the-loop doing the work it is for: not approving output the agent already trusts, but refusing to certify a silence as an answer.
Agent Capabilities
9 actionsRendering to PDF
2- Convert URL to PDF Renders a publicly reachable web page in a headless browser and stores the resulting PDF in FlowRunner file storage. PDFCrowd fetches the page server side, so the URL must use http:// or https:// and be reachable from the internet. Supports full page setup covering page size, orientation and margins, HTML headers and footers with automatic page numbering, viewport and scaling control for responsive layouts, custom CSS injection, and waiting for dynamic content before capture.
- Convert HTML to PDF Converts an HTML string into a PDF and stores it in FlowRunner file storage. This is the action for rendering invoices, contracts, quotes and certificates assembled earlier in the flow from a template. External assets referenced by absolute http:// or https:// URLs are downloaded during rendering, so images, web fonts and stylesheets must be publicly reachable. Supports page setup, headers and footers with page numbering, scaling and custom CSS.
Rendering to Image
2- Convert URL to Image Captures a full-page screenshot of a publicly reachable web page and stores it in FlowRunner file storage. The capture height matches the full document unless an explicit height is given. Supports PNG, JPEG, WebP, GIF, TIFF, BMP and ICO output, viewport width control, transparent or solid backgrounds, capturing a single element by CSS selector, and waiting for dynamic content. Use it for QA, compliance and monitoring records of a live page.
- Convert HTML to Image Renders an HTML string as an image and stores it in FlowRunner file storage. Useful for turning templated markup into social cards, email banners, chart snapshots or preview thumbnails. External assets referenced by absolute URLs are downloaded during rendering. Supports the full output format set, explicit width and height, transparent backgrounds and capturing a single element by CSS selector.
Reading Documents
3- Extract Text from PDF Extracts the text layer of a PDF while preserving the original reading layout, returning the text inline and also storing it as a .txt file. Use it to feed invoices, contracts or reports into downstream parsing, search indexing or AI steps. The PDF is fetched by PDFCrowd from a public http:// or https:// URL, and encrypted files can be opened with a password. Supports page ranges, layout-free extraction, paragraph detection, de-hyphenation and empty-line removal. This reads embedded text only and does not perform OCR on scanned images.
- Convert PDF to HTML Converts a PDF into an HTML document that reproduces the original layout, fonts and graphics, then stores it in FlowRunner file storage. Images, stylesheets and fonts are embedded into a single self-contained file. Use it to republish PDF collateral on the web or to make PDF content readable in a browser or an email client. Encrypted files can be opened with a password.
- Convert PDF to Image Rasterizes PDF pages into images and stores the result in FlowRunner file storage. Use it to build page thumbnails, document previews or images for OCR pipelines. A single-page result is stored as one image, while a multi-page result comes back as a ZIP archive with one image per page. Supports the full output format set, page ranges, DPI control, grayscale rendering and a forced ZIP archive for consistent handling.
Image Handling
2- Convert Image to PDF Places an image onto a PDF page and stores the result in FlowRunner file storage. Use it to turn scans, photos, signatures or generated charts into archival PDF documents. The image is fetched from a public http:// or https:// URL. Supports standard page sizes, orientation, margins, image placement, fit or stretch scaling, resizing, rotation and a page background color.
- Convert Image Format Converts an image between formats and optionally resizes, rotates or trims it, then stores the result in FlowRunner file storage. Use it to normalize uploaded assets, generate thumbnails, strip transparency or produce web-optimized WebP and JPEG variants. The source format is detected automatically and PNG, JPEG, WebP, GIF, TIFF, BMP and ICO output are all supported.
Frequently Asked Questions
What can FlowRunner do with PDF Export by PDFCrowd?
FlowRunner agents can run Convert URL to PDF, Convert HTML to PDF, and Convert URL to Image in PDF Export by PDFCrowd, plus 6 more actions.
Does connecting PDF Export by PDFCrowd to FlowRunner require OAuth?
No. PDF Export by PDFCrowd connects to FlowRunner with basic authentication (a username and password), no OAuth flow required.
Can PDF Export by PDFCrowd trigger a FlowRunner workflow automatically?
PDF Export by PDFCrowd doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with PDF Export by PDFCrowd
$100 in credits. No card required. Connect in minutes.