Why we built AI Card Vault — a working dealer's manifesto

I run a card store. Not as a side hustle, not as a marketplace operator, not as a VC-funded "platform." A real store. Ten thousand cards in active inventory. eBay, Whatnot, TCGPlayer, the occasional Shopify experiment. Pokemon-first because that's where the deals are right now, but sports and Magic and the rest come through too.

For two years I was a paying customer of every card-dealer tool you've heard of. CardDealerPro. CardLuma. CollX as a side channel. Spreadsheets, when nothing else fit. I bought scanners I didn't need. I learned PaperStream Capture settings I shouldn't have had to. I paid for tier upgrades I didn't want, to unlock features I felt I'd already paid for.

The thing I kept hitting, over and over, was this: the software didn't trust me to be a serious dealer.

It hid features behind paid tiers. It pushed me to a 30-minute Zoom demo before I could see how a button worked. It assumed I'd type out 80-character titles by hand for 487 cards in a row. It mangled my Japanese Pokemon because the catalog was built sports-first by someone who'd never sold an Eevee VMAX 269/SS-P.

So I built AI Card Vault. The product is real and the trial is real and the founder is the guy you'd Zoom with — except there's no Zoom call, because I'd rather you scan a card.

The premise

A vault is not a database. A vault is not a CRM. A vault is not a "platform for collectibles."

A vault is the place where serious things go to be kept correctly. You don't sleeve a 1999 1st Edition Charizard and then drop it in a shoebox. You don't grade a card and then forget which set it came from. The whole point of a vault is that the work has already been done — the card is identified, priced, catalogued, ready to go.

That's the product. Point your phone at a card. AI vision identifies it. We price it against PriceCharting plus live eBay sold comps. We hand you a draft listing that's ready to publish to eBay today — with TCGPlayer, Whatnot, CollX, and Shopify on the integration roadmap (no tier upgrade required when each ships). You review, you publish, you move on.

One card. One import, in our language. That's the unit of work. The tiers meter how many of those you can do per month. The Sleeve starts at $10. The Binder at $30. The Vault at $75. Every tier ships every feature and template, and every channel turns on the moment its integration ships. The only thing that scales is the import budget.

That last sentence is the entire pricing model in twelve words, and it took me a year of paying CardDealerPro to figure out why it mattered.

What the existing tools get wrong

I'll be direct about this because being indirect would be dishonest.

CardDealerPro gates features I'd already paid for. I paid $19/month for their Majors tier. I wanted to change the format of my eBay titles — a customer-facing string. I had to upgrade to a higher tier to do it. That's a tax on dealers who want their store to look like their store and not like a default template. It's also a textbook case of "feature gating the cheap thing." Custom Title Templates cost CDP roughly zero dollars to ship. Pricing power lives in the AI calls, not the string formatter.

I am not trying to "disrupt" CDP. The team there is competent and the product works. I am trying to articulate something they apparently don't believe — that every dealer who pays for the entry tier should get the entire feature surface. Meter the expensive thing. Don't paywall the cheap things.

CardLuma is eBay-only and English-only (as of this writing). I want my Japanese Pokemon cards titled in native script. I want the foreign-language Pokemon market — JP, KR, CN — covered properly. CardLuma can't do those things. Their pricing is honest — flat tier, all features — but the catalog is narrow.

Spreadsheets are free, until you count the hours. A 500-card breakers' box takes 20 to 30 hours of typing if you do it by hand. The same box runs through AI Card Vault in about 90 minutes, and that includes coffee.

COMC takes a cut, and you lose control. Consignment is fine if you don't want to be in the business. But you came here to be in the business.

What "Pokemon-first" actually means

It does not mean we don't do sports. We do. The catalog covers baseball, basketball, football, hockey, soccer, NPB, and the long tail.

It means we built the JP/KR/CN identification before we built sports edge cases. It means AI vision was tested on a 1996 Japanese Base Set Charizard before it was tested on a Topps Chrome Mike Trout. It means when you scan a Japanese-only Pokemon promo from a CoroCoro insert in 1998, the title comes back with the JP set code attached and a native-script option ready. Multi-region eBay (Germany today; Japan next on the integration roadmap) means the listing surfaces in the right markets, not just the US one.

I know how many Pokemon-first dealers are stuck on tools that ignore them. I am one of them. The product reflects that reality.

What "dealer-built" actually means

It means I sit in the same chair you sit in. Every feature ships because I needed it on a Saturday, not because a roadmap document said so.

The cert-import flow exists because I came back from a card show with eighteen PSA slabs and didn't want to hand-type the certificate numbers. I built it for myself on a Sunday. Now you can paste a list of cert numbers from any of the five majors — PSA, BGS, CGC, SGC, TAG — and the vault hits each grader's API, pulls the full record, and queues the listings.

The cross-channel conflict-resolution architecture exists because I've double-sold cards across manual workflows before — and I never want a single user of this software to do that. The logic is built and tested in the eBay path today; it activates per-channel as each new integration ships.

The bulk image-import-from-Google-Drive flow exists because shooting card photos with a phone burst is faster than running a $400 Ricoh autofeed, and I'd rather not buy hardware until I have to.

I could keep going. Almost every feature has a "the founder needed this" origin story. That's not marketing copy. That's the dev process.

What you should expect from us

Three things.

One: the trial is real and you don't pay to start. Fourteen days. Twenty-five free imports. No credit card. Scan a card in the first sixty seconds or the trial wasn't worth your time.

Two: we won't ever build a feature gate that puts a paywall between you and a custom title template. Every template, every custom field, every channel as it ships — on every tier. Whatever's true at the $75 Vault tier is true at the $10 Sleeve tier. The only thing that scales is your monthly import budget.

Three: when you write us, you're writing me. I read the support inbox. The founder is the support team for now. That'll change someday, but the day it changes is the day we hire someone who runs their own store.

What you should not expect from us

We are not a VC-funded platform play. We are not building a "marketplace for collectibles." We will not pivot to AI-generated card art. We will not, under any circumstances, send you a "we're excited to share!" email.

We will probably ship slower than CDP. We will definitely ship more carefully. The cost of a buggy bulk-listing tool is real money for the dealer who clicks Publish. I am writing this manifesto on a Tuesday and I will list 60 cards on Wednesday using the same software. If it breaks, I notice before you do.

Why "we're the vault"

A vault is a posture. A vault holds the thing you're not willing to lose. A vault is closed by default and opens on your terms.

CardDealerPro is a SaaS. CardLuma is a tool. COMC is a consignment lot. None of those words describe what serious dealers actually need. They need a place where the inventory lives. Where every card has been identified, priced, and queued correctly. Where the work is done before they pull a card to ship.

That place is the vault. We built it because nothing else existed that didn't make us pay extra or wait for a sales call or accept that our Japanese cards were second-class citizens.

The vault is open. The trial is real. The founder is in the inbox.

— Jamie

P.S. If you're a Pokemon dealer stuck in CDP's mid-tier paying for features you can't customize, or a CardLuma user who needs Japanese-language Pokemon coverage, or a spreadsheet operator with five thousand cards in a closet — start the trial. The first import lands in under sixty seconds.