From Any Document to Data You Can Put to Work
Scan from any scanner, use the RapidXtract app on your MFP, or drop digital files into a profile watch folder. RapidXtract extracts the information you define, applies your review rules, and prepares structured output for your business workflows.
Turn Documents Into Usable Information
Built for scanned and digital documents across your business, including tables and multi-page files.
Intelligent Extraction
Define the information you need, and RapidXtract extracts fields and line-item tables across varied document layouts, including multi-page files.
- Header fields and line items
- Cross-page table reconciliation
- Every value pinned to its spot on the page
Human-in-the-Loop Verification
A keyboard-driven QC workstation shows each value next to the highlighted source. Verifiers confirm, correct, capture text straight from the image and approve whole batches.
- Enter-to-confirm field flow
- Click or drag to capture from the image
- Row-by-row table confirmation
Data Ready for What's Next
Deliver extracted data as structured files to a destination folder, with the original document if you choose. Direct database export is also available when configured.
- JSON, CSV and optional original-file copy
- Configurable naming and folder rules
- Optional direct database export
However Documents Arrive, Keep Work Moving.
Scan from any scanner, use the MFP app, or drop digital files into a profile watch folder. RapidXtract takes it from there.
Capture
Scan documents or drop digital files into a profile watch folder to start processing.
Classify & Split
When configured, mixed batches are identified and separated into individual documents for the right profiles.
Extract & Enrich
AI extracts the fields and line items you define. Optional lookups can enrich the results with your reference data.
Review & Deliver
Results follow your profile's review rules before delivery as structured files or through configured database export.
Bring documents into the workflow in the way that suits your team:
Scan Your Way
Scan from any scanner into a configured watch folder, or use the RapidXtract app installed on your MFP.
Digital File Intake
Drop PDFs, TIFFs or image files into a profile watch folder for automatic processing.
Everything Your Document Operation Needs In One Platform
From document capture to usable output, configured around each document type and your review requirements.
AI Intelligence
- Vision-language model extraction
- Accurate reading of scanned documents
- Line items across page breaks
- Auto classification of mixed files
- Split by invoice / PO / DO number
- Self-hosted model or frontier model
- Per-profile model choice
- Processing at scale
Document Capture
- Scanner and MFP app intake
- Profile watch folders for digital files
- Barcode, QR & fixed-page splitting
- Orientation & skew correction
- Archive, exception & retry folders
- Insert missing pages during QC
- Batch naming rules
- Profile-specific intake rules
Security & Control
- Administrator & Verifier roles
- Per-profile verifier access
- Users, roles & profiles per customer
- Audit log of QC & admin actions
- Encrypted database passwords
- Page-credit usage control
- Session timeout protection
- Fully on-premise deployment
User Experience
- Keyboard-first verification
- Click-to-capture from the image
- Auto-fit zoom on every field
- Docked, resizable line-item table
- Live batch grid with filters
- Personal verifier dashboard
- Light & dark themes
- Real-time status updates
The Hard Parts, Handled
The details that decide whether automation actually saves time on the shop floor.
Automatically classify and split mixed documents
Drop a mixed batch into the Auto folder. RapidXtract identifies where each document begins, separates it from the batch and routes recognized document types to their matching profiles.
Documents without a matching profile are moved to an exception folder for review.
Fill gaps from your own database
Lookup rules run right after extraction: find the row in your SQL Server table where Invoice No matches, then fill PO Number, Vendor Code or anything else from that row. Fields filled this way are never guessed by the AI.
If a lookup can't run, the document goes to a verifier with the reason shown, and correcting the match value in QC re-runs the lookup instantly.
-- generated safely with parameters SELECT TOP 2 [PONumber], [VendorCode] FROM [dbo].[Invoices] WHERE [InvoiceNo] = @invoice_no AND [CompanyCode] = '1000' → PO Number = PO-2026-3101 (DB lookup) → Vendor Code = V-20417 (DB lookup)
Direct database delivery when you need it
Map profile fields to columns of a document table, and line items to a separate lines table, or let RapidXtract create the tables for you. Values are converted to each column's type, dates are read the way your region writes them, and re-exporting a document replaces its earlier rows in one transaction.
Pay only for pages that make it through
Page credits are allocated per project. Pages are reserved before a document is read and charged only when it is processed successfully. Failed documents release their pages automatically, and documents waiting for credit resume on their own once pages are added.
Wherever Documents Drive Work
RapidXtract is configured per document type, so the same platform serves very different paperwork.
Accounts Payable & Finance
- Supplier invoice capture
- Three-way matching inputs (invoice, PO, DO)
- Vendor and PO enrichment by lookup
- Structured data for finance workflows
Logistics & Freight
- Delivery orders and delivery notes
- Packing slips and packing lists
- Batch-scanned driver paperwork
- Per-depot scan stations
Manufacturing & Supply Chain
- Purchase orders from customers
- Goods-received paperwork
- Line items with codes and quantities
- Structured line-item data
Wholesale & Distribution
- High-volume invoice streams
- Mixed-document scanning via Auto folder
- Multi-page tables reconciled
- Exception handling for unknown papers
Shipping & Port Services
- Invoices with supporting dockets
- Service and vessel records kept with their invoice
- Document-number based splitting
- Auditable verification trail
Shared Services & BPO
- Many profiles, many teams
- Verifier access per profile
- Productivity per verifier
- Page usage per project
Learn RapidXtract Inside Out
Step-by-step guides for verifiers, administrators and IT teams in the Help Centre.
Quick Start Guide
From first sign-in to your first verified, exported document.
Read the guide →Verifier Handbook
Everything about the QC workstation: confirming, capturing, tables, lookups and batch actions.
Open the handbook →Profile Configuration
Watch folders, splitting, fields, lookups, AI model and output: every profile setting explained.
Configure profiles →Keyboard Shortcuts
The full shortcut list for fast, mouse-free verification.
See shortcuts →Deployment Guide
Docker deployment, environment settings, model server, database and upgrades.
Deploy RapidXtract →Troubleshooting
Answers to the most common questions about folders, jobs, exports and access.
Find answers →Ready to Stop Keying Documents by Hand?
See your own documents captured, processed and prepared for the systems and teams that need their data.
Frequently Asked Questions
What is RapidXtract?
RapidXtract is an intelligent document processing platform. Scan from any scanner, use the app on your MFP, or drop digital files into a profile watch folder. It extracts the fields and line items you define, follows your profile's review rules, and delivers structured files or exports directly to a database when configured.
Which file formats are supported?
PDF (scanned or digital), TIFF/TIF including multi-page files, JPEG/JPG, PNG and BMP. Mixed batches can be classified and split into individual documents when configured.
Does our data leave our network?
Not unless you choose so. RapidXtract runs on your own servers with Docker, and the AI model can be self-hosted on your own GPU server. Cloud-hosted frontier models are an optional, per-profile choice that uses your own API key.
How accurate is it, and what happens when it isn't sure?
Extracted values are linked to their source on the document so reviewers can check and correct them. Per profile, you decide whether documents always go to QC, go to QC when confidence is low, or continue without review. Database lookups that cannot run also send the document to a verifier.
Can it handle many documents scanned into one file?
Yes. Split by a fixed number of pages, at barcode or QR separator sheets, or wherever the invoice / PO / DO / DN / packing-slip number changes. Or drop the file into the Auto folder to have each document classified and routed to the right profile.
How is usage billed?
By pages. Page credits are allocated to projects; pages are reserved when processing starts and charged only when a document is processed successfully. Failed documents release their pages.