Snaplot
AI Cataloguing for Antiques Auctions
Why antiques cataloguing is hard
A 200-lot antiques sale is a research nightmare. Each lot is one-of-a-kind. You need to identify maker marks, date pieces, describe materials accurately, flag damage, and price against comparables. Six to twelve minutes per lot adds up to a full week of work per sale — and that’s before you factor in the inevitable “actually that’s a Coalport, not a Worcester” corrections.
How Snaplot helps
- Photo-grounded descriptions. The AI only describes what’s actually visible in your photos — no invented provenance, no fabricated maker attributions.
- Condition report from photo evidence. Chips, hairlines, repairs, areas of wear — flagged where the model can see them.
- Estimate range anchored to comparables. Snaplot looks up recent sold prices for similar items, then proposes a low–high range you approve or adjust.
- UK terminology by default. “Edwardian”, “Georgian”, “regency” — not American conversions.
- Confidence scoring. Each lot gets an AI self-rated confidence (1–10) so you know which need a quick human review and which are good to ship.
Where it slots in
| Job | Today (manual) | With Snaplot |
|---|---|---|
| Photograph the lot | You | You |
| Title | You research + write | AI drafts, you approve |
| Description | You research + write | AI drafts, you approve |
| Condition report | You handle, write up | AI from photos, you verify |
| Estimate | You research comparables | AI looks up comps, suggests range |
| Export to your auction platform | Manual reformatting | One-click XLSX |
Categories Snaplot handles well
- Furniture: Georgian, Victorian, Edwardian, Arts & Crafts, mid-century
- Ceramics: English porcelain (Worcester, Derby, Coalport, Wedgwood), Continental (Meissen, Sèvres), Asian export
- Silver: hallmarked English silver, Sheffield plate, Continental
- Glass: cut glass, art glass, paperweights, drinking vessels
- Pictures & prints: oils, watercolours, engravings (with provenance research aids)
- Clocks: longcase, bracket, carriage, mantel
- Decorative arts: Tiffany, Lalique, Doulton, Moorcroft
- Books, maps, ephemera
- Jewellery (with stone/metal identification from photos)
What Snaplot doesn’t try to do
Snaplot drafts. You decide. The AI is wrong sometimes — particularly on small marks that are blurred in photos, or on items so unusual that no comparable has ever sold. The confidence score tells you when to look harder. The output is a starting point, not a valuation.
Try it on your next sale
100 free lot credits, no card required. Run a 50-lot mini-test on your next sale to see whether the output lands. Most antiques rooms see 60–70% of lots ship straight from AI draft with minor edits, 25% need substantive editing, 5% need full re-write.
How does Snaplot handle obscure or one-off pieces?
It returns its best guess and flags low confidence. You add a few keywords (maker, period, material), re-run, and the AI grounds its description on those facts. Re-runs cost one extra credit per lot.
Can the AI identify maker marks?
For common UK and Continental marks, yes — particularly in well-lit close-up photos. Take a separate close-up shot of the mark; Snaplot uses every photo in the lot. For unusual marks, you’ll see “marks not legible” in the description and you can override.
More from Snaplot
Start cataloguing in minutes.
100 lots, no card required. Credits roll over. Cancel anytime.
Free 100 Lots Trial — Start Now →Questions? info@snaplot.co.uk