We Built the Engine Because
Nobody Else Would.
Xtracta is a Document Intelligence company headquartered in Auckland, New Zealand. We build the technology that turns complex, unstructured documents into clean, usable data. Our engine has been learning from real-world documents at production scale for over 14 years, and it powers document processing for over 1,000 businesses across every continent.
Where It Started
Before Xtracta existed, our founder Jonathan Spence was running IT at a scanning bureau. He was a customer of the document extraction software on the market at the time. It was slow, expensive, and required specialist engineers to build templates for every single document layout. Every new supplier, every new form, every new contract meant more templates, more configuration, more cost.
He thought: there has to be a better way. Software that learns from documents instead of being told what to look for. Software that gets better the more it runs, not worse. Software that any business can afford, not just enterprises with dedicated IT teams.
In 2010, he started building it. That engine is still running today. It has processed hundreds of millions of documents. It has learned from every single one. And it still doesn’t need templates.
What We Believe
Document Processing Is Not the Work
The work is the decisions, the momentum, and the scale that come after.
Every hour your team spends keying data from a document is an hour they’re not spending on the work that moves your business forward. We exist to make that hour disappear.
The Engine Should Do the Learning, Not Your Team
Set and forget. That’s the standard, not the aspiration.
Most tools need you to configure rules, build templates, and maintain them as documents change. Xtracta learns from the documents themselves. The longer it runs, the less anyone has to touch.
Good Technology Should Be Accessible
Touchless data capture should be a cost-effective option for organisations of any size.
When Jonathan started Xtracta, the tools on the market were built for enterprises with big budgets and dedicated teams. We built something that works for a business processing 500 invoices a month just as well as one processing 50,000.
Many Vendors Use It. We Built Around It.
Research-led since day one. Not retrofitted.
Xtracta has been performing research and development in this space for over 14 years. Long before it became a trend. Our research team holds PhDs, has published in international journals, and brings together academic rigour with commercial delivery.
Founded
Years of the engine learning
Pages processed monthly
Businesses worldwide
Team members
The Engine
At the centre of everything Xtracta does is the engine. It’s a research-led system that has been learning from real-world documents at production scale since 2013. Not test data. Not sample sets. Real invoices, real contracts, real bank statements from real businesses around the world.
The engine doesn’t rely on templates or rules. It reads layout, language, context, and structure. It understands what type of document it’s looking at, finds the data that matters, and knows what it means. When it encounters a format it hasn’t seen before, it applies what it has learned from over a decade of production data.
In 2024, the team rolled out production-ready deep learning transformer models for the most common document types. The engine keeps evolving. But the principle hasn’t changed since day one: set it up, let it learn, and forget about it.
Our Research Commitment
Xtracta has been performing research and development in this space long before it became a technology trend. Our research team is a diverse group of specialists who each bring deep expertise in their field. Several hold PhDs. Many have published in international journals. They bring together academic thinking with commercial discipline to turn ideas into products that work in production.
Xtracta is a recipient of high-tech funding from the New Zealand government for its research programme. The team has over 55 years of collective experience in research and development, and over 200 years of collective experience in technology and software.
55+
Years collective R&D experience
200+
Years collective tech experience
PhDs
On the research team
NZ Govt
High-tech funding recipient
The Team
We’re a team of about 30 people across Auckland, Australia, and Vietnam. Engineers, data scientists, researchers, and client advocates. Small enough that every customer relationship matters. Experienced enough that the technology works at global scale.
Jonathan Spence
Founder & CEO
Graham Hill
Director
Darren Tuit
Research Team Lead
Bishal Gauli
Customer Services & Implementation
The Journey So Far
2010
Jonathan Spence founds Xtracta.
The mission: build document extraction software that learns, doesn’t need templates, and is affordable for any business.
2013
The first version of the engine launches after several years of research and development.
2014
The API goes live.
Any software, cloud or on-premise, can connect, submit documents, and retrieve data. White-labelling launches so partners can brand Xtracta as their own.
2015
Rapid growth as multiple ERP platforms embed the technology.
Auto-provisioning tools launch. Geo-distributed processing nodes go live around the world.
2016
Mobile app launches for receipt capture.
Xtracta hits 100%+ compounded sales growth.
2018
Major system re-architecture for container-based, scalable infrastructure.
New models for complex line item extraction.
2019
Batch processing system released.
Multiple major enterprise clients deploy Xtracta for business-critical processing.
2020
Private cloud and on-premise deployment options launch for customers requiring enhanced security or data sovereignty.
Strong growth continues through the pandemic.
2022
Research begins on deep learning transformer approaches.
The primary goal: much-improved out-of-the-box extraction across all document types.
2024
First production-ready deep learning transformer models rolled out
Xtracta becomes a certified Peppol access point for eInvoicing.
Now
Over 1,000 businesses worldwide.
10 million+ pages processed monthly. The engine keeps learning. The team keeps building.
Technology and Integration Partners
Xtracta runs on world-class infrastructure and integrates with the most widely used business software on the planet.
FAQ
Xtracta is a Document Intelligence company headquartered in Auckland, New Zealand. We build the technology that turns complex, unstructured documents into clean, usable data. The engine has been learning from real-world documents at production scale for over 14 years, and it powers document processing for over 1,000 businesses worldwide.
Xtracta was founded in 2010 by Jonathan Spence. The first version of the engine launched in 2013 after several years of research and development. The company has been growing continuously since, with 100%+ compounded sales growth in 2016 and 2017.
Xtracta is headquartered in Auckland, New Zealand, with team members in Australia and Vietnam. The platform runs on geo-distributed infrastructure with regional data centres around the world, powered by AWS, Microsoft Azure, and Google Cloud.
About 30 people. Engineers, data scientists, researchers, and client advocates. The research team includes specialists with PhDs and international journal publications. Collectively, the team has over 200 years of technology experience and over 55 years of R&D experience.
Xtracta has been building research-led document intelligence technology since 2010, long before it became a trend. The engine learns from real-world documents, not templates. It doesn’t need specialist engineers to configure. And it’s designed to be affordable for organisations of any size, not just enterprises
Learn more: How Xtracta Works · For Business Owners · For Software Companies · For Partners
See What Xtracta Can Do for Your Business.
This isn’t a self-service sandbox. Our team sets it up with you. Your data, from day one.