What we do
We process handwritten documents — single pages, decades of letters, batch archives, business forms — and return text you can edit, search, and export. We work with a wide range of customers, from individuals digitising family archives to businesses processing hundreds of thousands of pages a month, and we’re constantly improving both the underlying models and the workflow around them.
Behind the scenes, we run AI models trained on hundreds of millions of public-domain documents to deliver accuracy traditional OCR can’t match — in a scalable, resilient system built for high-volume processing without compromising privacy.
Who we are
Handwriting OCR is run by a small team headquartered in central London, United Kingdom. The same people who build the product handle customer support — when you write in, you’re getting someone who actually knows the system, not a queue of generic agents.
We’re independent, fully UK-based, and operate under UK data protection law. For anyone handing over sensitive material — family archives, medical records, legal correspondence — where a service is run from genuinely matters, and we welcome the accountability that comes with it.
How we started
Handwriting OCR began with a personal problem. After his grandfather died, founder Matthew Wrigley spent months trying to digitise decades of letters and notebooks by hand. The work was slow and made one thing obvious: traditional OCR had never been built for handwriting, and so many important documents — family histories, archives, business records, field notes — were sitting in drawers because nobody could face the transcription.
By late 2023, modern AI had finally made it possible to read the kinds of documents traditional OCR couldn’t handle at all. Handwriting OCR was started to focus on exactly that: accuracy on the hardest documents, in a service that just works.
Our stance on your data
A lot of AI companies train on whatever you feed them. We don’t.
Documents you upload are used to provide the service you’ve asked for — and nothing else.
- Never used for training. Not our models, not anyone else’s.
- Never sold or shared with third parties.
- Auto-deleted on the schedule you set in your account.
That isn’t a marketing line; it’s a hard constraint we run the company on. If you’re handing over family letters or sensitive business documents, you should expect nothing less.
Stories we love to hear
One of the things we love most about this work is hearing from the people who use it. Readers who’ve unlocked precious memories from a grandparent’s letters. Businesses that have revolutionised their admin operations. Researchers who’ve finally been able to search a decades-old archive.
If you’ve got a story to tell, drop us a line — we read every one.
Or you can try it for yourself — free, no card required.