CoinMarketCap Crypto Scraper

Cryptocurrency data from CoinMarketCap: the ranked, paginated listing table (price, market cap, volume, % change across timeframes) plus rich single-coin profiles (description, official links, all-time high/low, market cap dominance).

What you get

Two independent inputs, either or both:

listingQueries — one SEARCH_SUMMARY + N CRYPTOCURRENCY rows per entry. Each coin row: id · name · symbol · slug · cmcRank · tags · price · marketCap · volume24h ·

percentChange1h/24h/7d/ 30d/60d/90d/1y

coinSlugs — one CRYPTOCURRENCY row per slug, a richer profile: description, urls (official website/explorer/source-code/social links), all-time-high/low with dates, 52-week range, marketCapDominance, holders, watchCount. This is a genuinely different, larger response shape than a listing row — see "Field shapes differ" below.

Input

{
  "listingQueries": [
    { "cryptoType": "all", "sortBy": "market_cap", "sortType": "desc", "convert": "USD", "limit": 100 }
  ],
  "coinSlugs": ["bitcoin", "ethereum"]
}

All listingQueries fields are optional (the defaults shown above apply). coinSlugs are CoinMarketCap URL slugs — the last segment of https://coinmarketcap.com/currencies/<slug>/.

sortBy and convert are validated lists, not free text

sortBy must be one of market_cap, price, volume_24h, percent_change_24h, name, circulating_supply — each individually verified to genuinely reorder results. An unrecognised sortBy is not honest on this API: it silently answers a clean, empty result (totalCount: 0) rather than an error, so only verified values are accepted. convert must be one of USD, EUR, GBP, JPY, IDR, BTC, ETH — a short, verified set rather than the full list of currencies CoinMarketCap itself supports, which wasn't exhaustively checked this session.

cryptoType (all / coins / tokens) is a genuine filter — coins excludes stablecoins and wrapped assets, verified by comparing top results.

Field shapes differ between the two record sources

A listing row and a detail row are not forced into one shape. Most notably, both carry a tags field, but it means something different depending on which endpoint answered: a plain list of strings on a listing row (["mineable", "pow", ...]), a list of {slug, name, category, priority} objects on a detail row. This is raw passthrough of what CoinMarketCap's own API actually returns from each endpoint, not an inconsistency introduced by this actor — see CRAWLING_METHOD.md for the full reasoning.

Known limits

1. Bogus lookups are reported honestly, but upstream disguises the error. A coinSlugs entry that doesn't exist gets a clean ERROR row with _error: "not_found" — but CoinMarketCap's own API answers that case with an HTTP 200 wrapping

"error_message": "The system is busy, please try again later!"

2. No search-by-name. Only exact slug lookup is offered for coinSlugs — CoinMarketCap's own keyword-search endpoint (a different path than the data API used here) wasn't explored this session. Use listingQueries with a high limit and inspect the slug field to find the right one first.

3. convert only offers 7 currencies. CoinMarketCap supports far more (dozens of fiat currencies), but only these 7 were individually verified this session — see "sortBy and convert are validated lists" above.

Anti-bot posture

None encountered. 8/8 TLS profiles clean, and a 10-request burst with zero delay between calls came back 200 every time during testing — the opposite of properties/99co-properties-scraper, built immediately before this actor the same session, where an equally small burst triggered a genuine Cloudflare challenge. Paced modestly by default anyway (0.3s minimum interval) as routine good behaviour, not because any throttling was observed.

Policy

https://api.coinmarketcap.com/robots.txt (User-agent: *) has no ClaudeBot/anthropic-ai/GPTBot named group, and disallows only /v1/ and /v2/ — this actor uses /data-api/v3/, covered by neither line.