Anyone who has traded cards at a store or convention knows the scene: two piles on the table, a phone out, and five minutes of searching card by card while the other person waits. Both sides want a fair trade, but nobody wants to do the maths. A trade calculator does it in one step: you enter what each side is offering, and it shows the total for each pile and the gap between them.

This post describes one such tool, who it is for, how it works, which endpoints it needs and how many requests a day a busy evening takes.

Who it is for

The tool fits three kinds of users.

  • A game store that hosts weekly trade nights and wants a tablet at the counter so staff can settle disputes quickly.
  • A community app or Discord server where members arrange trades and want a neutral number to start from.
  • A convention seller who trades at a booth and needs to answer "what would you give me for this?" without guessing.

It works for any game in the catalog, and a single trade can mix games, for example a Pokemon card against two Magic cards.

How it works

The tool has four parts: a card picker, a price lookup, a fairness rule and a result screen.

1. A card picker

People need a way to say which card they mean. Use the search endpoint and let them pick from the results with the picture.

GET /v1/cards?game=Pokemon&name=mega rayquaza&limit=10

Each result has card_id, name, set, number, rarity, game and image_url. Showing the picture and the number avoids mix-ups between printings. If your users scan cards with their phone camera, the card scanner endpoint (POST /v1/cards/match) turns the name and number an AI model reads off a photo into the matching card, so the picker can be skipped entirely. The card scanner guide shows how.

Typical chase cards that show up on trade tables right now:

Mega Rayquaza ex - 138/103ME06: Delta Reign$286.98Near Mint HolofoilCaptain James T. Kirk (Showcase) (Surge Foil)Star Trek$88.22Near Mint Foil

2. One lookup for the whole trade

Once both sides have picked their cards, send all the card_id values in one request, with the condition the pile is in.

POST /v1/cards/lookup
{"card_ids": ["...", "..."], "condition": "Near Mint"}

The answer lists each card with its SKUs. For a trade you usually read three numbers per card:

Field Use in the tool
market_price The default value of the card in the trade
listing_price_low The cheapest copy for sale right now, a sanity check
sold_price_low and sold_price_high The range of recent sales, to show how firm the price is

Cards that are not found come back in not_found, so you can ask the user to pick them again instead of silently counting them as zero. If a card has several printings (normal and foil, for instance), the SKU's printing field tells you which price belongs to which version, and the picker should ask the user to choose.

Piles in different conditions need a lookup per condition. Many trade-night pairs agree on "Near Mint or close", so one condition setting at the top of the screen is usually enough, with a switch for "Lightly Played" on the cards that need it.

3. A fairness rule

Add up each side and compare. Then let the organiser set the rules once:

  • A tolerance, such as "within 5% counts as even".
  • A haircut for cards under a minimum value, such as "cards under a dollar count as zero".
  • Which price to trust: market_price for a neutral view, or listing_price_low plus shipping when the trade is between people who would otherwise buy online.

Because values are passed on exactly as reported, the numbers on your screen match what the API sends, which keeps arguments short. Prices can be null for a card with no recent data, so show "no price" for those lines and leave them out of the total.

4. The result screen

Show the two totals, the difference and a plain verdict ("Side B is $12.40 short"). Add the card pictures so both people can check the list. A button to add cash to even the trade is a small feature that gets used a lot.

Market prices and recent sales refresh every 6 hours, so a trade on Saturday afternoon uses numbers from that day, not from last month's spreadsheet.

5. Checking a doubtful card

When one side says "that card just spiked", the tool can show the last month of sales for that card:

GET /v1/cards/{card_id}/history?timeframe=30d

The answer includes daily market_price, sale_price_low, sale_price_high and quantity_sold. A flat line with plenty of sales settles the argument. A card that sells twice a month deserves a wider tolerance. Thirty days is available on Starter and up; Free includes 7 days, and Pro and up go back a full year.

How many requests a day it needs

Take a store with a busy weekly trade night: 40 trades in an evening, with about 10 cards on each side.

Step Requests
One price lookup per trade, 20 cards 40
History check on 1 in 4 trades 10
Card searches, about 6 per trade 240

That is around 290 requests and about 800 cards priced on one evening. A single Starter lookup can carry up to 20 cards, which is exactly one trade, and Starter allows 1,000 requests and 1,000 cards a day. Short history (7 or 30 days) counts as one request, while 3-month, 6-month and 1-year history counts one request for each card.

For a community app with many users trading at the same time, Pro allows 6,000 requests a day and 100 cards per request, and Business allows 25,000 requests a day with 200 cards per request. You can see the plan's limits and what you have used with the free GET /v1/usage call. A request over a limit gets a 429 with a Retry-After header, so the app can show a short "try again in a moment" message instead of failing.

The Free plan (100 requests a day, one card per request) is enough to build and test the screens with a few friends. It is for personal projects.

What you could build

  • A trade-night tablet app for a game store, with the store's own rules for tolerance and cash top-ups.
  • A Discord command that takes two card lists and replies with both totals and the difference.
  • A convention booth tool that shows a buy price and a trade price side by side for every card a visitor puts on the counter.
  • A trade history log for a community, so members can see how fair past trades were and which cards moved most.
  • A mixed-game trade screen for stores that stock Pokemon, Magic, One Piece and more, using one lookup call for all of them.
  • An AI agent that answers "is this trade fair?" through the MCP server, using the same key.

Every paid plan, from Starter up, allows commercial use, so a store or app that charges for the tool can run on it. Prices come from TCGplayer, and TcgPriceApi is independent and not affiliated with TCGplayer.