One app writes a beautiful description of your antique. The other tells you what someone will actually pay for it.
If you’ve been weighing antique identification apps, Curio is probably on your shortlist. It’s the longest-running app in this corner of the App Store — launched in July 2024 by Dionysus Labs — and it sits at 4.8 stars across roughly 12,000 ratings. Reviewers love the polish. The descriptions read like museum placards. The collection view is gorgeous.
We make Circa, so this isn’t a neutral review. What we’ll try to do is be specific about where the two apps actually differ, because they don’t differ where most comparisons assume they do. Both identify antiques from a photo. Both build a collection. The interesting gap is in how each one decides what your item is worth.
Key Takeaways
• Curio is best-in-class for descriptive narrative and collection design; Circa is built around live market data and conversational refinement.
• Multiple Curio reviewers — including a working antique-shop owner — report that pricing draws from asking prices, not sold prices, leading to inflated estimates.
• Circa pulls live comparable listings from Google Shopping and auctions, then runs them through a separate AI pricing assessor that returns a confidence score.
• Curio offers no chat or follow-up refinement; if the AI gets it wrong, you scan again.
What Curio does well
Curio is a beautifully made app. The first time you scan something with it, you understand why people give it five stars. The identification page reads like a short essay — rich in historical context, written with the kind of confident voice you’d expect from a print catalog. One reviewer put it perfectly: it “gives so much detail it almost writes the for-sale ad for you.”
The collection management is the other obvious strength. Items are organized into folders, the typography is restrained, and the recently added cloud sync (which arrived in v1.8.1 in November 2025) means you don’t lose your catalog when you switch phones. PDF export rounds it out. If you’re documenting a personal collection for the pleasure of having a documented personal collection, Curio is genuinely well made.
It’s also widely available — iOS and Android both, with over 100,000 Android downloads and roughly 12,000 ratings on the iOS side. For an app that’s under two years old, that’s real traction.
Where Curio falls short
The complaint that surfaces most often in the negative reviews isn’t about identification. It’s about pricing. One review — from someone who identifies as an antique-shop owner — lays it out plainly: “It rarely gets the prices correct. It seems to return results based on a small sampling of asking prices rather than a large sampling of sold prices.”
That’s a structural problem, not a bug. Asking prices and sold prices are different numbers — often by a lot. Anyone who’s ever browsed eBay knows the gap. A Victorian sideboard listed at $3,200 might sell for $1,400. A signed pocket watch listed at $800 might quietly close at $310. If a pricing engine is reading the listing prices and not the sale prices, it’s reading the wishful thinking of sellers, not the actual market.
Curio also struggles in the categories where pricing matters most. Reviewers single out silver, watches, and coins as weak spots — categories where the difference between a generic identification and a maker-specific one moves the price by an order of magnitude. And a few users have flagged that the same item, scanned twice from slightly different angles, can return materially different identifications.
The other gap is what happens when the AI gets it wrong. Curio doesn’t have a chat feature. There’s no way to add a photo of the maker’s mark, type in a detail you noticed, or correct the model. You can rescan, but you can’t refine. For a single-photo identification flow, that’s a meaningful constraint.
Try a different approach.
Sourced pricing, chat refinement, and a confidence score on every estimate.
Download Circa for iOSHow Circa approaches it differently
Circa was built around a different starting point: the price needs to be sourced, the identification needs to be refinable, and the user should see what the AI is looking at.
When you scan an item, two AI models run in parallel — a fast pass for an immediate identification, and a deeper pass that arrives within ten to fifteen seconds. The pricing pipeline is separate from the identification one. Once Circa knows what your item is, it queries live Google Shopping listings and auction databases through Serper, gathers a real basket of comparable items, and feeds those results to a dedicated AI pricing assessor. That assessor reads each comparable, weighs condition, filters out the irrelevant matches, and returns a price range with an explicit confidence score and a short reasoning paragraph.
You can see the comparables. They’re shown to you with source attribution — Google Shopping, auction result, dealer listing — and a link out. If the AI is working from three relevant matches, you know it. If it’s working from twelve, you know that too.
The chat layer is where Circa diverges most clearly from Curio. After any scan, you can ask follow-up questions, attach more photos, type in details, or request a deeper analysis. Found a maker’s mark on the bottom of a vase? Snap it. The identification updates and the price range moves with it. That conversation is what most identifications actually need — the first photo rarely tells the whole story.
Side-by-side comparison
Two well-built apps with overlapping audiences. The differences are sharper than they look from the App Store screenshots.
| Feature | Circa | Curio |
|---|---|---|
| AI identification | Multi-model (fast + deep) | Single-pass |
| Pricing source | Live Google Shopping + auctions | Not disclosed; reportedly asking-price based |
| Confidence scores | Yes, on every estimate | No |
| Comparables shown with source | Yes | Visual matches, source not surfaced |
| Deep Appraisal pass | Yes, with reasoning | No |
| AI chat for refinement | Yes, with photo support | No |
| Currencies | 14 supported | USD primary |
| Cloud sync across devices | No (on-device only) | Yes (added Nov 2025) |
| PDF catalog export | No | Yes |
| Descriptive narrative | Concise, fact-led | Long, essay-style |
| Platform | iOS | iOS & Android |
Pricing compared
Both apps are subscription-based and both run an aggressive paywall after a free-scan limit. The structure differs.
Curio. Three free scans, then a paywall. Subscription tiers shift over time, but the App Store currently shows in-app purchases ranging from a $6.99 weekly tier up to $49.99 for some annual plans. Curio appears to test pricing across multiple SKUs, so what you see in the paywall depends on the cohort you land in. Free trial isn’t explicitly disclosed.
Circa. Three free scans, then yearly at $49.99 with a 7-day free trial, or weekly at $4.99. The trial unlocks the full app — chat, Deep Appraisal, unlimited scans, multi-currency, full collection. No tier shuffling.
The pricing here is roughly comparable. The difference is what each subscription is paying for. Curio buys you the polished description and the cataloging. Circa buys you the description, the cataloging, the live market data, and the chat refinement layer.
Which app should you choose?
Choose Curio if the pleasure of antique collecting is mostly the storytelling for you. You want something that reads like a museum and looks like a coffee-table book. You scan items mostly out of curiosity, you don’t need the price to be defensible, and you appreciate a polished feed of finds. Curio is also the natural choice if you’re on Android.
Choose Circa if you ever need the pricing to be right. That includes resellers, dealers, anyone valuing an estate, anyone documenting items for insurance, and collectors who’d rather see five real listings of comparable items than a single confidently-stated number with no source. The chat layer is the other reason: most identifications get better with one follow-up question, and Circa is the only app in this comparison that lets you ask one.
The bottom line
Curio is what an antique app looks like when it’s designed for the joy of identifying things. Circa is what one looks like when it’s designed for the moments where the answer matters.
If you’re curious about both, try Circa first. Three scans are free. You’ll see the comparables, the confidence score, and the chat in the first few minutes — and you’ll know whether the trade-off is worth it for the way you actually collect.