--- title: "Docparser Integration" description: "Extract structured fields from PDFs with Docparser. Agents upload documents to a parser by URL, poll processing status, and retrieve parsed invoice, receipt, and purchase order data as JSON to route into accounting or CRM systems." url: https://flowrunner.ai/integrations/docparser date_modified: 2026-08-01T02:40:32-07:00 --- # Docparser [Document Processing](https://flowrunner.ai/integrations/category/documents-forms) Extract structured fields from PDFs with Docparser. Agents upload documents to a parser by URL, poll processing status, and retrieve parsed invoice, receipt, and purchase order data as JSON to route into accounting or CRM systems. [Verified](https://flowrunner.ai/integrations/verified "What does verified mean?") · 7 actions · Basic auth · available [Docparser website](https://docparser.com/) · [Platform Documentation](https://docparser.com/) · Capability data verified 2026-07-27 1. A vendor invoice PDF arrives in the accounts payable mailbox and lands in cloud storage 2. Upload Document submits the file URL to the invoice parser, which returns a document id 3. Get Document Status polls until import and processing both complete 4. Get Parsed Results retrieves the extracted fields: invoice number, vendor, dates, line items, total 5. The agent validates the extraction: totals sum, vendor matches a known record, no duplicate invoice number 6. Anything that fails validation goes to the AP clerk with the parsed fields shown next to the source PDF 7. Confirmed records flow into the accounting system, and the clean majority never needed a human at all ## What This Integration Enables Docparser turns the worst kind of business input, the PDF, into the best kind, structured JSON. You define a parser in the Docparser web app for each document type, invoices, receipts, purchase orders, and the parser extracts the same named fields from every document that matches its layout. This connector wires that engine into FlowRunner: agents upload documents by URL, poll asynchronous processing, and pull extracted fields individually or in bulk, then carry them into the accounting, CRM, and spreadsheet systems where the data actually earns its keep. The honest framing: parsing is probabilistic. A vendor redesigns their invoice template, a scan comes through crooked, a line-item table gains a column. The teams that win with document extraction are the ones who build the validation and review path on day one, which is precisely what FlowRunner's [human-in-the-loop](https://flowrunner.ai/concepts/human-in-the-loop) steps exist for. - Extract invoice, receipt, and purchase order fields from incoming PDFs, images, and Word files - Route parsed records into accounting and CRM systems without retyping - Sync newly processed documents on a schedule with date-filtered result listing - Correlate parsed documents to your own records with a remote id - Route different document types to different parsers, each with layouts per vendor variation ### Without FlowRunner **Data entry as a job description**: Someone reads PDFs and retypes numbers into the accounting system, all day, with error rates to match **Documents queue behind a person**: Invoices wait in a mailbox until the one person who does intake gets to them **Every document gets equal attention**: The clean invoice and the garbled scan cost the same human minutes ### With FlowRunner **Documents become records on arrival**: PDFs upload, parse, and land as structured JSON within minutes of hitting the mailbox **Intake runs around the clock**: Parsing and validation continue overnight and on weekends, no queue owner required **Attention goes only where it earns**: Validated extractions flow straight through; only the failures and anomalies see a person ## Use Case Scenarios ### The AP inbox that processes itself Invoices arrive by email and save to cloud storage. For each new file, the agent calls Upload Document against the invoice parser, polls Get Document Status until processing completes, and fetches the fields with Get Parsed Results. Valid records append to the payables tracker in [Google Sheets](https://flowrunner.ai/integrations/google-sheets) and post to the finance channel in [Slack](https://flowrunner.ai/integrations/slack) with the key fields and a link to the source PDF. The AP clerk's job shifts from typing invoices to ruling on exceptions. ### Purchase orders into the CRM before the sales call ends Customers send purchase orders as PDFs. The agent uploads each to the PO parser, waits out processing, and pushes the extracted record, customer, PO number, items, amounts, into [Airtable](https://flowrunner.ai/integrations/airtable) keyed by a remote id that ties it back to the originating deal. When sales asks "did their PO come through and does it match the quote," the answer is a record, not a search through attachments. ### The nightly sync that catches what real-time missed Some documents arrive through channels no trigger watches. Every night, the agent calls List Results in Processed After mode with yesterday's cutoff, pulling every document the parser handled in the last day regardless of how it got there. New records reconcile against the tracking system; anything parsed but never routed gets flagged. The result is a guarantee the per-document flows cannot make alone: nothing that Docparser processed ever silently disappears. ## Human-in-Loop Highlight Get Parsed Results hands back numbers that are about to become money. In an AP workflow the extracted total flows toward a payment run, and paying a mis-parsed invoice is a real loss: a transposed digit in the total, a duplicate invoice slipping through under a slightly different number, a vendor name matched to the wrong account. None of that is reversed by an API; it is reversed by phone calls. So FlowRunner agents score every extraction before it moves: totals must sum from the parsed line items, the vendor must resolve to a known record, and the invoice number must be new. Documents that pass flow through untouched. Documents that fail render side by side, parsed fields against the source PDF, for the AP clerk, whose correction both fixes the record and tells the team which parser layout needs attention. The agent does the reading. A person stands between a bad parse and a bank transfer. Agent processes routinely Detects exception requiring judgment Clear match Continues automatically Ambiguous Routes to human via preferred channel Human decides Agent resumes with decision ## Agent Capabilities 7 actions ### Account 1 - **Test Connection** Verifies the configured API key by pinging the API, returning a pong response on success. The first step of any setup or troubleshooting flow. ### Parsers 2 - **List Parsers** Lists every parser configured in the account with id and label. Each parser defines the extraction layout for a document type; the id addresses uploads and result fetches. - **List Parser Models** Lists the model layouts defined for a specific parser. Multiple layouts let one parser handle document variations, such as invoices from different vendors. ### Documents 2 - **Upload Document** Submits a document to a parser by publicly accessible URL. Processing is asynchronous; the returned document id is used to poll status and fetch results. Supports PDF, images, and Word documents. - **Get Document Status** Retrieves a document's processing status: import progress, processing state, format support, page count, and timestamps. Poll until both import and processing complete. ### Results 2 - **Get Parsed Results** Retrieves the extracted structured data for one processed document, keyed by the parser layout's field names. Returns nested objects or a flattened record. The output every downstream system consumes. - **List Results** Lists parsed data for a parser's recently processed documents, filterable by upload or processing date and by remote id. The Processed After mode is built for periodic syncing of new results. ## Frequently Asked Questions ### What can FlowRunner do with Docparser? FlowRunner agents can run Test Connection, List Parsers, and List Parser Models in Docparser, plus 4 more actions. ### Does connecting Docparser to FlowRunner require OAuth? No. Docparser connects to FlowRunner with basic authentication (a username and password), no OAuth flow required. ### Can Docparser trigger a FlowRunner workflow automatically? Docparser doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger. **Work at Docparser?** This integration exposes Docparser to AI agents on every FlowRunner plan, including through MCP, at no cost to you. [See what FlowRunner offers integration partners](https://flowrunner.ai/integrations/partners), including how to keep this page current. --- Markdown version of https://flowrunner.ai/integrations/docparser. Site index: https://flowrunner.ai/llms.txt