Final Fantasy, Sorcery, Gundam & Riftbound Added to Our Recognition System

Until now, our trading card game recognition system has covered the largest TCGs, including Pokémon, Magic: The Gathering, Yu-Gi-Oh!, and Flesh and Blood. These models are already integrated into some of the biggest marketplaces and apps worldwide, automating routine tasks like cataloging, scanning, and trading – freeing collectors to focus on the fun parts.

Like the games themselves, our card database is constantly evolving and expanding. We try to cover as many games as possible with high accuracy and quality, ensuring your website or app meets the highest standards. This update is available for testing via the Ximilar App and demo, and for integration through our REST API.

What is the AI Card Identifier?

Ximilar develops AI recognition systems for trading cards, sports cards, and other collectibles. Our technology is used in collector websites, mobile apps, and trading platforms to:

• Identify cards instantly from a photo
• Provide structured catalog data
• Retrieve listing and sales price history
• Promote listings on marketplaces such as eBay, Rakuten, or Racuma

Key Attributes Provided by the TCG Identifier

• Short and full card name, card number within the set
• Set and set code, year of release
• Language/alphabet (e.g. English, Japanese, Chinese)
• Foil/holo, side and rarity
• Pricing data and links to marketplaces

The services are available through an API, interactive demos, and the Ximilar App. For a full overview, see the service page or the API documentation. The quickest way to explore the solutions is to book a call to discuss your specific use case.

ONE API TO RULE THEM ALL

AI That Understands Collectibles

Recognize, analyze & grade collectibles, find them in databases and marketplaces. Our API was built by collectors for collectors.

Explore & Try it now

Covered Games

Our recognition system now covers a wide range of trading card games:

• All-time classics, such as Pokémon, Yu-Gi-Oh!, and Magic: The Gathering
• Modern hits, including One Piece, Flesh and Blood, Digimon, or Dragon Ball Super
• Anime & crossovers like Weiss Schwarz, Union Arena, Vanguard, and now Gundam
• Fantasy & indie favorites, e.g., Force of Will, Grand Archive, and now Sorcery & Riftbound
• Newer & special games, for instance, MetaZoo, Lorcana, Star Wars Unlimited, Marvel Champions, or Garbage Pail Kids

So let’s test the newly added games via demo, Ximilar App and REST API to see how they will perform.

Final Fantasy Trading Card Game

Final Fantasy was globally loved long before its TCG launched. The Final Fantasy Trading Card Game was introduced in Japan in 2011 and expanded globally in 2016. With over 20 million booster packs shipped worldwide, frequent expansions, and an active organized play scene, the game attracts more and more new players every year. New sets typically include around 150 cards, drawing characters from the games and movies, like Sephiroth, Terra, or Lightning.

In this example, I tried the recognition of Final Fantasy trading cards in a low-quality binder image via our card ID demo. The largest detected card was identified as Balthier Final Fantasy XII #5-156S, with details including card side, rotation, grading, ID, number, element, full and short name, rarity, set and set code, and a link to eBay.

Sorcery: Contested Realm

My favorite new addition is Sorcery: Contested Realm TCG. Created by Erik Olofsson and Nickolas Reynolds and launched via Kickstarter in 2022, it features hand-painted artworks and deep strategic gameplay. Its small but passionate community is quickly growing among collectors who value both playability and the beautiful artworks.

In this example, I uploaded a staged promo image from Sorcery’s website, showing a small, tilted card, to Ximilar Demo. The system accurately detected the card’s position and identified it as Screaming Skull, providing details such as grading status, short and full name, release date, rarity, set slug, type, year, and a link to eBay – all in just 1.08 seconds.

Gundam

Since its debut in 1979, Gundam has grown into one of the most influential mecha franchises in anime. The Gundam Card Game, released in summer 2025, brings its iconic mechs and pilots into TCG form. With multilingual editions and a global fanbase, the game is poised for rapid growth.

For testing, I used a user-generated image of a graded 001 Aile Strike Gundam card in a labelled slab. The system accurately identified the game and card details – including card ID, side, number, full and short name, set with code, and a link to eBay.

The graded slab label was detected as a separate object, with information such as grading company (PSA), grade (10), card number, certificate number, verbal grade, set, year, and a link to psacard.com for certificate verification.

Riftbound: League of Legends TCG

Another addition this year is Riot Games’ Riftbound, which brings the League of Legends universe – including the hugely popular Arcane series – into physical cards. Thanks to its connection to one of the world’s largest gaming communities, Riftbound is gaining momentum fast.

For this example, I tested an Annie Fiery card via Ximilar App. The test form can be found under Collectibles Recognition: TCG Identification. The system provided Annie Fiery’s category, side, set, full and short name, year, card number, card ID, subcategory, and a link to eBay listings.

Improvements in Pokémon TCG Identification

Our recognition system now includes advanced support for identifying “first generation” prints of the Pokémon TCG – those sets where the artwork stays the same, but subtle design details differ between releases. By analyzing minute features such as holo-foil stamps, set symbols, typography, and card stock variations, our AI can accurately distinguish between First Edition (Limited Edition) or Base Set, Unlimited, Shadowless, and Legendary Collection cards (read more here).

Let’s test the upgraded trading card game recognition system on a true legend – the 1999 Base Set Charizard Shadowless Holo (#4), graded by PSA as 8 Nearly-Mint and currently listed on eBay for $39,999.99.

And here’s how the system dealt with the original Japanese Charizard 006 from 1996:

We also expanded the language support. Our system now fully recognizes both Simplified Chinese (Mainland China) and Traditional Chinese cards. We look forward to supporting Korean and other languages soon, so stay tuned!

How to Test the Solutions via API

Whenever you test a card in the app, the cURL command and API response appear right under the results table. Here’s the request from my example:

curl --url https://api.ximilar.com/collectibles/v2/card_id \
  --request POST \
  --header "Content-Type: application/json" \
  --header "Authorization: Token YOUR_API_TOKEN" \
  --data '{
    "records": [
      {
        "_url": "https://artiholics.com/wp-content/uploads/2025/07/Screenshot-2025-07-01-161622.png"
      }
    ],
    "pricing": false
  }'

Here is a selected excerpt from the full API response. As you can see, the response includes even more details than what is shown in the results table in the App, for instance, alternatives to the card.

{
  "records": [
    {
      "_url": "https://artiholics.com/wp-content/uploads/2025/07/Screenshot-2025-07-01-161622.png",
      "_status": { "code": 200, "text": "OK" },
      "Category": "Card/Trading Card Game",
      "_tags_simple": ["Card/Trading Card Game", "front", "Riftbound", "rotation_ok"],
      "_identification": {
        "best_match": {
          "name": "Annie Fiery",
          "year": 2025,
          "card_number": "001",
          "set": "Proving Grounds",
          "set_code": "OGS",
          "card_id": "OGS-001",
          "full_name": "Annie Fiery Proving Grounds #OGS-001",
          "subcategory": "Riftbound",
          "links": {
            "ebay.com": "https://www.ebay.com/sch/i.html?_nkw=Riftbound+Annie+Fiery+OGS+%23001&_sacat=2536"
          }
        },
        "alternatives": [
          { "name": "Caitlyn Patrolling", "set": "Origin", "card_number": "068", "card_id": "OGN-068" },
          { "name": "Lux Crownguard", "set": "Proving Grounds", "card_number": "014", "card_id": "OGS-014" },
          { "name": "Leona Determined", "set": "Origin", "card_number": "238", "card_id": "OGN-238" },
          { "name": "Teemo Strategist", "set": "Origin", "card_number": "121a", "card_id": "OGN-121a" }
        ]
      },
      "Graded Slab": [{ "name": "no", "prob": 0.98155 }]
    }
  ],
  "status": { "code": 200, "text": "OK" }
}

A detailed how-to, including quickstart and first steps, is outlined in the API documentation, which I recommend consulting when needed.

Each API call requires your unique user authentication token and a few setup steps in the App before calling the API. All of this is explained in the first steps guide on our blog:

Get Started: First Steps Guide

Stay Tuned for More

We remain dedicated to enhancing our already supported games. There’s always room for improvement, and we are striving for 100% accuracy for every card you scan.

If you’re looking to integrate card identification into your app, website, or platform, or need guidance on using the API, don’t hesitate to reach out – our team is ready to help you get started and ensure the solution fits your needs.