TCGplayer Price Monitor

Monitors TCGplayer market price and listed median price for trading-card products (Pokemon, Magic: The Gathering, Yu-Gi-Oh!, Lorcana, One Piece, and more) — one row per product x printing (Normal / Foil / 1st Edition ...), with optional run-over-run change detection so you only react to prices that actually moved.

Pure JSON-over-HTTP (no browser), built for cheap scheduled monitoring runs.

Quickstart (3 lines)

1. Enter one or more search queries (e.g. charizard ex obsidian flames) or paste known product IDs / product URLs (e.g. 610532 or https://www.tcgplayer.com/product/610532).
2. Run the Actor — each product yields one dataset row per printing with marketPrice and listedMedianPrice.
3. Schedule it (e.g. daily) with a fixed productIds list: with enableChangeDetection on, every row carries priceChanged, prevMarketPrice, and marketPriceDelta versus your previous run.

What it does

• Resolves search queries via TCGplayer's product search (paginated, capped by maxResultsPerQuery).
• Accepts explicit productIds (skips search; metadata is resolved in one batched lookup) — the reliable mode for scheduled monitoring.
• Fetches per-printing price points for every unique product: marketPrice, listedMedianPrice, buylistMarketPrice.
• Joins in search metadata: set name, rarity, card number, lowest listed price, listing count.
• Change detection: keeps a snapshot of your last run in a named key-value store (tcgplayer-price-monitor-state) and emits priceChanged / prevMarketPrice / prevListedMedianPrice / marketPriceDelta / changedAt.
• Contract test on every run: a known-good product's response shape is validated; on upstream schema drift the run warns loudly (status message + one uncharged warning row) instead of silently emitting garbage.

What it does NOT do

• No seller-level listings or seller-filtered pages (those are disallowed by tcgplayer.com robots.txt — we do not touch them).
• No buying, sniping, or cart automation.
• No historical backfill — price history accrues from your own scheduled runs.
• No fabricated values: if TCGplayer has no market price for a printing, you get null (a price of 0 is also treated as "no price").

Input examples

One-shot search:

{
    "queries": ["charizard ex obsidian flames"],
    "productLine": "pokemon",
    "maxResultsPerQuery": 25
}

Scheduled daily monitor of a fixed list (recommended):

{
    "productIds": ["610532", "113670", "https://www.tcgplayer.com/product/477236"],
    "printings": ["Foil"],
    "enableChangeDetection": true
}

Output example (real rows from a test run, 2026-08-29)

{
    "productId": 610532,
    "productName": "Teal Mask Ogerpon ex - 177/131",
    "setName": "SV: Prismatic Evolutions",
    "productLine": "Pokemon",
    "rarity": "Hyper Rare",
    "number": "177/131",
    "printingType": "Foil",
    "marketPrice": 8.79,
    "listedMedianPrice": 9.86,
    "buylistMarketPrice": null,
    "lowestPrice": 6.78,
    "listingCount": 189,
    "sealed": false,
    "url": "https://www.tcgplayer.com/product/610532",
    "scrapedAt": "2026-08-29T14:32:23.077Z",
    "prevMarketPrice": null,
    "prevListedMedianPrice": null,
    "marketPriceDelta": null,
    "priceChanged": false,
    "changedAt": null
}

On a later run, after the market price moved:

{
    "productId": 610532,
    "printingType": "Foil",
    "marketPrice": 8.79,
    "prevMarketPrice": 7.50,
    "marketPriceDelta": 1.29,
    "priceChanged": true,
    "changedAt": "2026-08-29T14:33:16.634Z"
}

(second example abbreviated; all rows always contain the full field set)

Notes on fields:

• priceChanged is true when marketPrice or listedMedianPrice differs from the previous snapshot for that product+printing.
• On the very first run there is no previous snapshot, so all rows have priceChanged: false and prev*: null.
• productTypeName is emitted for completeness but TCGplayer's search payload currently does not populate it; use the sealed boolean to separate sealed products from singles.

Pricing (pay per event)

Event	Price	Charged when
Run start	$0.001	once per run
Product price row	$0.003	for every product x printing row delivered to the dataset
Price change row	$0.004 (additional)	only for delivered rows with priceChanged: true

Failed fetches, filtered-out printings, nonexistent/unknown product IDs, and the schema-drift warning row are never charged.

Worked example: monitoring 500 rows daily (a row = one product x printing; use the printings filter to get exactly one row per product) where ~10% move per day costs about 500 x $0.003 + 50 x $0.004 + $0.001 = **$1.70/day ($51/month)**. A daily 100-row watchlist under the same assumptions is $0.34/day ($10/month).

Scheduling recipe

1. Create an Apify Schedule (e.g. daily at 09:00) pointing at this Actor with your fixed productIds input.
2. Add an Actor integration/webhook on run success and consume the dataset filtered to priceChanged = true (the dataset API supports field filtering), or plug the dataset into Make/Zapier/Google Sheets.
3. The change snapshot lives in the named key-value store tcgplayer-price-monitor-state in your account — delete it to reset change detection.

Data source & respect

Data comes from TCGplayer's public, unauthenticated JSON endpoints (the same ones the website itself uses) over plain HTTP — no browser, no login, no CAPTCHA circumvention. The Actor is deliberately polite: low concurrency, spacing between requests, exponential backoff on 429/5xx, and a descriptive User-Agent identifying this Actor. It never crawls HTML pages and never touches seller-filtered pages, which tcgplayer.com's robots.txt disallows. Prices are public market data; no personal information is collected.

Limitations

• These are unofficial endpoints; TCGplayer can change them without notice. A schema contract test runs on every run and flags drift immediately; we target same-day fixes, 14 days worst case.
• Search relevance is TCGplayer's — broad queries return whatever the site's search ranks first (may include sealed products; filter with sealed/productLine).
• lowestPrice / listingCount come from the search payload and are null for products whose metadata lookup failed.
• If more than 30% of requests to an endpoint fail, the run aborts with a clear error rather than emitting partial garbage.

Support

Use the Issues tab on this Actor's page — reports are answered within 24 hours, and breaking upstream changes are fixed within 14 days (usually much faster).