PackRips API

Get every number on PackRips as JSON. It's free and needs no key.

33 sites46 packs with pulls countedCORS open

Endpoints

GET/api/operatorsEvery site, ranked, with average return, typical pull, pulls counted, grade, sell-back rate and source.

GET/api/operators/{site}One site: the same figures plus payout terms, returns by game and all its packs.

GET/api/packsPacks from every site. Filter by operator, category, currency, minN or listed. Sort by rtp, n, price or median.

GET/api/packs/{site}/{pack}One pack: returns for today, this week and all time, published odds next to what the pulls show, the spread of pull values and recent pulls.

GET/api/pullsRecent pulls, up to 100 per page (the newest 200 per pack from the last 14 days). Filter by operator, category or minMultiple. Sort by when, value or multiple.

SPEC/api/openapi.jsonOpenAPI 3.1 description of all of the above.

Every endpoint returns JSON. A bad parameter gets a 400, and an unknown site or pack gets a 404.

Quick start

Rip sites ranked by average return:

curl -s https://packrips.org/api/operators | jq '.operators[] | {name, rank, rtp: .stat.rtp, n: .stat.n, source: .stat.source}'

The ranked Pokémon packs (500 or more pulls counted, the boards' threshold), best first:

curl -s 'https://packrips.org/api/packs?category=pokemon&minN=500&sort=rtp&limit=10'

One site, then one of its packs with its 50 newest pulls:

curl -s https://packrips.org/api/operators/courtyard
curl -s 'https://packrips.org/api/packs/courtyard/PACK-SLUG?pulls=50'

Pulls worth 10 times the pack price or more, biggest first:

curl -s 'https://packrips.org/api/pulls?minMultiple=10&sort=value&limit=20'

Parameters

/api/pulls

operator: a site slug, for example courtyard
category: pokemon, one-piece, magic, basketball, football, baseball, soccer or sports
minMultiple: only pulls worth at least this many times the pack price
sort: when (the default), value or multiple. Page with page, and set limit up to 100.

/api/packs

Takes operator, category, currency (cash, crypto or store_credit), minN, listed (true, false or all), sort, order, page and limit up to 500. The full schema is in openapi.json.

The meta block

Every response starts with meta, so a number never travels without its pulls counted and its source:

{
  "meta": {
    "publisher": "PackRips (https://packrips.org)",
    "docs": "https://packrips.org/api",
    "license": {
      "name": "CC BY 4.0",
      "url": "https://creativecommons.org/licenses/by/4.0/",
      "note": "Applies to every figure; all are measured by PackRips. Payout terms text is seeded from RipRanks' published operator data."
    },
    "source": "feed",
    "provenance": "Measured by PackRips",
    "n": 760,
    "computedAt": "2026-10-13T15:09:46.605Z",
    "generatedAt": "2026-10-16T01:47:06.011Z"
  }
}

That's the live meta for the whole site right now. source is feed for every figure: PackRips measures them all from the sites' live feeds. n is the number of pulls counted behind the figures in the response.

Each stat inside carries the same fields (rtp, median, n, source, computedAt). rtp is the average return: all card value ÷ all money spent, before any sell back. median is the typical pull as a share of the pack price. The methodology explains every field.

Agent tools (WebMCP)

AI agents in your browser can call these 5 read-only tools on any PackRips page.

list_operatorsList card-rip sites

get_operator_statsGet one rip site's stats

list_best_packsRank packs by observed return

get_packGet one pack's stats

get_recent_pullsGet recent pulls

PackRips declares them through WebMCP (document.modelContext.registerTool, with a fallback for browsers that shipped it as navigator.modelContext). In other browsers nothing happens. Each tool calls the endpoints above and returns their JSON.

Every tool is marked readOnlyHint. Tools that return pulls also set untrustedContentHint, because card and pack names come from the sites' own feeds. The WebMCP draft defines no discovery file, so PackRips publishes none, and agents find the tools on the page.

Using the dataEvery figure is PackRips' own measurement, licensed CC BY 4.0: use them for anything and credit "PackRips (packrips.org)" with a link. Payout terms text is seeded from RipRanks' published operator data and isn't PackRips' to license.Please cache responses. The edge caches them for 60 seconds (15 seconds for pulls) and returns are worked out again every five minutes, so polling faster than once a minute gets you the same answer. There's no hard rate limit today.The API has no version number yet. New fields may appear, and anything that breaks existing clients will be announced on this page first.Questions or higher limits: hello@packrips.org

For language models

/llms.txt describes the site, the data and this API in a few hundred words. /llms-full.txt adds the full methodology and the current rankings in one file.