CardIntel

The price guide that doesn't lie to you.

Every card. Every market. One API.

Collectors get the flagged comp, the evidence and the band. Builders get the same rows over one API.

Open the appGet an API key
Charizard VMAX
pokemonswsh4#020/185
Rare Holo VMAX
Card Intel fair value
$26.80
p25 to p75
23.50 - 31.00
Confidence
0.85
Comps
47 / 90d
  • tcgplayer$28.50 / 31 obs
  • cardmarketno data
  • 30d trend-3.2%
GET/v2/cards/{card_id}
{  "id": "3f7c1b90-2ad5-4f0e-9c2b-8ad1b70f4e11",  "game": "pokemon", "name": "Charizard VMAX",  "set_code": "swsh4", "number": "020/185",  "prices": {    "tcgplayer": { "median_usd": 28.5, "observation_count": 31 },    "cardmarket": null,    "cardintel_fair_value": {      "median_usd": 26.8, "p25_usd": 23.5, "p75_usd": 31.0,      "confidence": 0.85, "observation_count": 47, "window_days": 90    }  },  "trend_30d_pct": -3.2}
Documented v2 example response. Same fact, both renderings. The empty venue is printed, not hidden.
223,906
TCG cards indexed
66,194
Sports card definitions
1,862,844
Source observations
2 / 18
Venues live / registered

Two venues. One card. 53x apart.

That is one real card in our own report window. Most guides average a spread like that into a single clean line and never show you it existed.

Observed cross-venue average sale prices for PSA 10 cards, May 16 to July 22 2026
Card, PSA 10Spread
Caitlin ClarkPanini Prizm #27ebay $159/fanatics $8,500$8,34153.4x
Shohei OhtaniTopps Chrome Update HMT1ebay $759/fanatics $8,252$7,49310.9x
Shohei OhtaniBowman Chrome #71ebay $588/fanatics $6,896$6,30811.7x
Victor WembanyamaPanini Prizm #136ebay $1,312/fanatics $6,600$5,2885.0x
Larry BirdTopps Chrome #8 Interstellarebay $168/fanatics $4,269$4,10125.4x

Market Transparency Report #1. 65,117 completed sales totalling $25,935,941 across two live venues over a 67 day window. Averages are per venue, not blended.

What we do with it
Both venue averages stay on the card, with the observation count behind each one. We report the spread instead of averaging it into a clean line.
What a flag carries
Cited sale IDs, the time window, a confidence band, and the model version that raised it. Six detectors run over the comp stream.
What a flag is not
Never a named party and never a finding of fact. Every flag has a rebuttal window, and one dismissed under review cannot be re-raised without fresh evidence.
Our own coverage gap

In that same window, 12.2% of ingested listings matched a canonical card definition. The other 87.8% were excluded from every aggregate rather than force matched into one.

Read the methodology

Odds first. Then the box math.

Sealed EV is only as honest as the hit rates underneath it. We start from observed print runs, not from a marketing sheet.

box EV=Σ(hit rate×fair value)vssealed ask→rip or hold
hit ratein the API
Printed odds and observed print runs today. Community break aggregations are the next source tier.
fair valuein the API
The p25 to p75 band with its confidence score, already served on every card by /v2.
sealed asknot yet
Sealed SKU matching runs on the existing eBay and Fanatics lanes. Not ingested yet.

Until the third input is real, we do not print an EV number. That is the same rule every other figure on this page runs on.

Observed hit-class scarcity, most flagged releases
Player
Rare
Total
Ratio
  • Elly De La CruzUSC1502024 Topps Chrome Update1,6205,3873.3x
  • Babe Ruth#32025 Topps Chrome Baseball2,1066,3483.0x
  • Reggie Jackson#1522025 Topps Chrome Baseball1,9675,6602.9x
  • Roberto Clemente#212025 Topps Chrome Baseball2,0025,6952.8x
  • Victor Wembanyama#1362023 Panini Prizm4169142.2x

Rare parallel to total print run, from the same report window. A ratio running above the documented distribution is what raises a pull-rate flag. It is a pattern, not a finding of fact.

Eight games, one schema, one key.

REST over JSON with an OpenAPI 3.1 schema. Cards, sets, price history and fair value across eight TCGs, plus sports sold comps.

Get an API keyRead the docs
Request
curl -H "X-Card-Intel-Api-Key: $CARDINTEL_KEY" \
  "https://api.cardintel.gg/v2/cards?game=pokemon&q=charizard&limit=20"
Response20060 req/min · 1,000 req/day on the free key
{
  "data": [
    { "id": "3f7c1b90-...", "name": "Charizard VMAX",
      "set_code": "swsh4", "number": "020/185",
      "rarity": "Rare Holo VMAX" }
  ],
  "total": 37, "limit": 20, "offset": 0
}
Every endpoint, every tier
  • GET /v2/games
  • GET /v2/cards
  • GET /v2/cards/{id}
  • GET /v2/cards/{id}/prices
  • GET /v2/games/{game}/sets
  • GET /v2/market/movers
  • GET /v2/me
  • GET /v2/health
Games covered
  • Pokémon
  • Magic: The Gathering
  • Yu-Gi-Oh!
  • One Piece
  • Disney Lorcana
  • Flesh and Blood
  • Digimon
  • Star Wars Unlimited

Plus sports sold comps and active listings from eBay and Fanatics Collect, on the same spine the app reads.

How keys work
Header
X-Card-Intel-Api-Key
Introspection
GET /v2/me returns your tier and enforced limits
Open route
GET /v2/health takes no key and no budget
Over the limit
429, counted per key across the whole /v2 surface

Free keys are issued by hand while self serve signup is being built. Tell us what you are building.

The two rows nobody else fills in.

Card Ladder flags suspicious sales and publishes a confidence level too — credit where it is due. What no one else lets you do is argue back with a logged rebuttal, or see the venues they are missing.

Feature and pricing comparison between Card Ladder, Market Movers, GoCollect and Card Intel
CapabilityCard LadderMarket MoversGoCollectCard Intel
Typical price$0 to $20/mo$10 to $50/mo$0 to $9/mo$0 to $39/mo
Cross-venue sold compsyes, 21 marketplacesyessold comps by grade2 live, 18 registered
Suspicious compsflagged, published dailynot surfacednot surfacedsurfaced with evidence
Confidence on the number1 to 5, by sale recencynone publishednone published0 to 1, p25 to p75 band
Dispute a flagno pathno pathno pathrebuttal window, logged
Public APInonoyesyes, free key
Publishes its own coverage gapsnononoyes, transparency report

Competitor rows reflect publicly documented behaviour as of August 2026. If a row here is wrong, tell us and it gets corrected on this page first.

One spine. Two ways in.

The app and the API read the same rows out of the same database. Pick the surface, not the dataset.

The app

for collectors

Every sold comp plotted, the suspicious ones flagged with evidence, and a band you can inspect.

  • Free
    Look up a card, see the basics.
    $0
    forever
    live
  • Collector
    The honest price guide.
    $14
    / month
    planned
  • Pro / Flipper
    For people who buy to sell.
    $39
    / month
    planned

Annual billing is decided at roughly two months free, and applies once consumer checkout opens.

Open the app

The API

for builders

The same rows as REST, keyed per developer so your quota is yours and not your network's.

  • Free
    60 req/min
    $0
    forever
    live
  • Builder
    600 req/min
    TBA
    planned
  • Scale
    6,000 req/min
    TBA
    planned

Rate limits above are the values the limiter enforces today, not a target.

Get an API key
Public beta

Nothing on this page can charge a card today. The Developer API is live and rate limited per key. The app is read only, and saved watchlists, price alerts and portfolio tools are planned rather than shipped. See the full pricing breakdown for what each tier includes.

One list for both SKUs
Ask for early accesshello@card-intel.com

Founder pricing when consumer billing opens, free keys for early builders. There is no signup form yet, so this reaches a person.

Nothing here is smoothed in secret.

Probabilistic, not accusatory

A flag is a statistical statement about a pattern. Participant identities are stripped server side and no output names anyone.

Banded and disputable

Every number carries a confidence score. Every flag carries a rebuttal window and an append only dispute log.

Research, not advice

Card Intel shows fair value against ask and which comps look wrong. It does not issue buy or sell verdicts.

Open the appGet an API key