Skip to content
Sortup

Solution · Document AI

Invoices, estimates and forms turned into clean data your systems accept.

For logistics, construction and finance teams retyping data from PDFs.

How it works

Document extraction, end to end.

  • Already yours
  • We build
  • Human step
Document extraction: the system in 5 steps
  1. 01

    Documents

    PDFs in any layout

  2. 02

    Extraction

    Vision and language models

  3. 03

    Review

    Uncertain fields checked

  4. 04

    Retraining

    Corrections improve the model

  5. 05

    Your systems

    JSON, ERP or database

Capabilities

Five jobs it handles.

  • 01Any PDF layout, vendor by vendor
  • 02Fields to your JSON schema
  • 03Review on fields the model is unsure of
  • 04Fine-tuning from your own corrections
  • 05Output your ERP or database accepts

Shipped work

Work you can open.

Built by Sortup or one of our co-founders, as labelled.

All client work
  • More shipped work9client case studiesAll client work

FAQ

Questions about this one.

Every way to work with us
Do we need training data?

A sample of your real documents is enough to start. Corrections made during review become training data.

Can it handle every vendor’s layout?

Yes. Where layouts differ a lot, we fine-tune a model per vendor.

Where does our data go?

It stays in your accounts wherever possible, and model providers are used under API terms that exclude training on it.

Start with a 14‑day Agent Pilot?

No pitch deck. An engineer, and a straight answer.

Hours
We work yours, in the GCC, the UK and EU or the US
NDA
We sign yours before the first call
IP
You own the code from the first commit