Photograph the lot. Approve the entry.
Dataforge Cataloguer drafts the title, description, condition, category and quantity of a lot from a few phone photos in about 15 seconds. Your team checks it, edits it and sends it straight to your auction system.
A 20-minute call, on your own lots if you send photos ahead.
Every lot waits for someone to type it up.
A lot can't go into a sale until it has a title, a description, a condition and a category. When cataloguing falls behind, so does everything after it.
- 01
Nothing sells until it’s catalogued
Slow cataloguing means slower sales and a fuller warehouse. Goods-in keeps arriving whether the descriptions are written or not.
- 02
Skilled people, typing
Experienced cataloguers spend their day typing up items they understood at a glance.
- 03
Three ways to describe one item
Every cataloguer writes differently, so the same item gets described three ways in one catalogue.
- 04
Everything handled twice
Photos are taken on the warehouse floor and the entry is keyed in later, at a desk.
From photos to a catalogue entry in four steps.
Your cataloguer takes the photos and approves the entry. The typing in between is done for them, and nothing is sent until they say so.
- 01
Photograph
A few angles, plus any model plate or label, on any phone.
- 02
AI drafts
Title, description, condition, category and quantity, in about 15 seconds.
- 03
Review
Edit anything, then approve. Anything it wasn’t sure about is highlighted.
- 04
Send
Approved entries go straight into your auction or stock system.
What the AI does, and what it won't.
It writes the first draft. Your team approves every entry, edits anything they want and keeps the final say on what goes in the catalogue.
What it does
- Writes the title and description
- Auction-style, in your house style, from the photos.
- Picks a category from your list
- Your own categories, not a generic set.
- Judges condition
- A grade and a short note, from what the photos show.
- Counts items
- So a pallet of 14 power tools is listed as 14.
- Reads serials and model plates
- When they are clearly visible in a photo.
- Picks the lead photo
- The best shot goes first.
- Flags what needs a human look
- Asset tags, possible safety issues and anything it isn’t sure of.
What it won't do
- Invent a serial number or spec
- If it can’t see it, the field is left blank.
- Publish or send without approval
- Every entry is approved by a person first.
- Decide whether an item should be sold
- That stays with your team.
Built for the warehouse floor.
It runs on the phones your team already carries. If the signal drops at the back of the warehouse, stalled items retry on their own.
No app store
It’s a web app that installs to the home screen. Sign in once and stay signed in.
Choose the description style
Short, standard, bullet list or full, set before the AI runs.
Recovers by itself
Stalled items retry automatically. A failed draft gives a “try again” button, and the photos are kept.
Team view
Supervisors see every lot in progress, who catalogued it and what was changed.
Send us photos of a few typical lots and we'll show you the drafts on the call.
Book a demoIt sends entries into the system you already run.
It's connected to a major ERP and stock system in production at a UK commercial auction house. We connect to other auction and inventory platforms through their API.
The entry arrives through the system's API, against the right vendor and goods-in record.
Checked before it’s sent
The reviewer sees which vendor and goods-in record an entry is going to, so a wrong match is caught before it reaches your system.
Your categories
Entries use your own category list and house style, so they arrive ready for the sale.
Through the API
Tell us what you use and we’ll confirm on the call how we’d connect to it.
Your photos and catalogue data stay yours.
- UK and EU hosting
- The app, the database and the photos are hosted in the UK and EU.
- No training on your data
- Enterprise AI terms: your photos and catalogue data are not used to train AI models.
- Your own accounts
- Each person signs in with their own account, with cataloguer and admin roles.
- Every action logged
- Who created, edited and approved each entry, and when.
- 09:41M. PriceM. Price: Photographed lot 1048, 4 photos
- 09:41AI draftAI draft: Entry drafted in the standard style
- 09:43M. PriceM. Price: Changed quantity from 12 to 14
- 09:44M. PriceM. Price: Approved the entry
- 09:44SystemSystem: Sent to the stock system, goods-in GI-20928
When someone asks who changed a description or who approved a lot, the answer is in its history.
Start with a pilot on your own lots.
It runs next to the way you catalogue now until your team is happy with it. How long set up takes depends on your system, so we'll talk it through on the call.
Book the demo call- 01
Demo call
20 minutes, on your own example lots if you send photos ahead.
- 02
Set up
We tune the AI to your past catalogues and categories, and connect your system.
- 03
Side by side
Your cataloguers use it next to the current process while we tune it.
- 04
Switch over
It becomes the normal way of working.
Stratas has helped more than 150 organisations automate document‑heavy work.
More from DataforgeNo. It takes the typing away. Your team still checks and approves every entry, and they always have the final say.
The cataloguer sees the draft before anything is sent and can edit any field. Fields it isn’t sure about are highlighted, and it leaves details blank rather than guess.
Yes. We set it up using examples from your past catalogues and your own category list.
It’s connected to a major ERP and stock system in production, and we can connect to other auction and inventory platforms through their API. Tell us what you use and we’ll confirm on the call.
A phone with a camera and an internet connection. No scanners or special hardware.
Priced per house, based on volume. We’ll quote after the demo.
See it catalogue your lots.
Send us a few photos of a typical lot and we'll show you the draft entries on the call.
A 20-minute call. No commitment.