If you've ever tried to price a K-pop photocard, you know the pain: the same card is $8 on one app, $45 on a reseller's Depop, and "DM for price" everywhere else. There's a $2B+ collectibles market here with almost no price transparency.
Every price-guide app and reseller tool I've looked at solves this the same way: scrape eBay and Mercari. Which makes sense if you're in the US. But it misses something big — the cards originate and price-discover in Korea first. A card trades on the Korean market for weeks before the Western resale price settles. If you're only watching eBay, you're reading yesterday's news.
So I went to the source.
The source market nobody's scraping
Pocamarket (포카마켓) is Korea's dedicated photocard marketplace — 477,000+ cards listed, organized by group, member, era, and card type (album / fansign / POB / lucky-draw). It's where Korean collectors buy, sell, and set the reference price. For a huge fraction of cards, this is the primary market and eBay is the derivative one.
The interesting part for anyone building tooling: it has a clean public catalog API. No login needed for the listing data.
GET https://pocamarket.com/apis/card/gb/v2/search?q=&page=
Returns per-card, structured: group, member, price, trade type (sale vs. trade), stock, wish count, sales volume, and an image. Everything you need to build a price index — per member, per era, per card type — sourced from where the price actually forms.
What you can do with it
Arbitrage signal. Korean price vs. eBay/Mercary price = the spread. When a card is cheap on Pocamarket and hot on Western resale, that's a sourcing signal before the Western price catches up.
A real price index. sales_volume + wish_count per card let you build a demand-weighted index per member/era — the "StockX for photocards" that collectors keep asking for but nobody's built on source-market data.
Inventory pricing. If you resell, pricing against the origin market instead of guessing beats the DM-for-price economy.
A couple of honest caveats
Prices are the catalog's, not seller identities. The scraper I built collects card + price data only — no seller nicknames, IDs, or profiles. (Both a privacy rule and a legal one — don't scrape people.)
card_type and era need the detail endpoint, not just search. The base catalog gives you the price signal; the per-card detail call adds the metadata.
Sort is fixed to newest-first on the public API — server-side price/popularity sort params are ignored, so you sort client-side.
If you don't want to maintain the scraper
I packaged this as a ready-to-run Apify actor — it handles the pagination, the detail enrichment, and the privacy filtering, priced per result:
https://apify.com/kdatafactory/pocamarket-scraper
(Want the retail side too — K-pop album/goods prices in USD from Ktown4u's global store? There's a matching actor: https://apify.com/kdatafactory/ktown4u-scraper. Origin-market secondary prices + origin-market retail prices is the full picture.)
But the endpoint above is the real unlock. Whether you build it yourself or not, the point stands: if you're pricing K-pop collectibles off Western marketplaces only, you're trading on lagged data. The source market is right there, and it's structured.
Happy to answer questions on the data shape or the index math in the comments.
Top comments (0)