nike-research

Nike research

Browse and search Nike's catalog and pull product detail, colorways, reviews, and nearby store locations — all as normalized JSON from the Crawlora API, with no HTML scraping.

When to use this skill

• "What does X cost on Nike?" or "find X on Nike."
• "Browse this Nike category" / "what's in Nike's [Men's Running Shoes]?"
• "Compare colorways/prices for this Nike shoe."
• "Pull the reviews for this Nike product."
• "Find the nearest Nike store to [location]."

Setup (one-time)

• Get a free Crawlora API key (2,000 credits/mo, no card) at https://crawlora.net.
• export CRAWLORA_API_KEY=sk_your_key_here
• All requests: x-api-key: $CRAWLORA_API_KEY against https://api.crawlora.net/api/v1. Missing/invalid key → 401.

How it works

1. Browse categories — /nike/categories lists Nike's Men/Women/Kids/Jordan taxonomy; each subcategory carries a slug for browsing.
2. Search or browse — /nike/search takes exactly one of keyword or category (never both). keyword runs Nike's own relevance search; category browses a taxonomy slug (from /nike/categories or a prior response's facet_nav). Both return product groups with pricing, colorway images, and every purchasable color variant, paginated with page. Use /nike/suggest (query) first if you want typeahead ideas for a vague keyword.
3. Product detail — /nike/product fetches one color variant, keyed by the pair slug + style_color pulled from a search result's colors[].slug / colors[].style_color — there's no single product id.
4. Reviews — /nike/product/reviews (same slug + style_color pair, paginated with page) returns written reviews plus an aggregate rating breakdown.
5. Nearby stores — /nike/stores (lat, lng, optional radius_miles) returns nearby physical stores with address, phone, and distance.
6. Compare the JSON fields (price, colorway, rating, distance) across items or stores and answer.

Full endpoint list, methods, and params: reference/endpoints.md.

Calling the API

# Browse a category (slug from /nike/categories):
scripts/crawlora.sh /nike/search category="mens-shoes" | jq '.'

# Keyword search:
scripts/crawlora.sh /nike/search keyword="air max" page=2 | jq '.'

# Product detail (slug + style_color from a search result's colors[]):
scripts/crawlora.sh /nike/product slug="nike-air-max-90-mens-shoes-6n3vKB" style_color="CN8490-002" | jq '{title,price}'

# Reviews:
scripts/crawlora.sh /nike/product/reviews slug="nike-air-max-90-mens-shoes-6n3vKB" style_color="CN8490-002" | jq '.'

Raw curl fallback:

curl -fsS -H "x-api-key: $CRAWLORA_API_KEY" \
  "https://api.crawlora.net/api/v1/nike/search?keyword=air+max" | jq '.'

Endpoint reference

See reference/endpoints.md for every Nike endpoint this skill uses (method, path, params, description).

Examples

• Category sweep with colorway comparison: /nike/categories to find a subcategory slug, then /nike/search with category set, paginated, to list products and compare price/colorway across each result's colors[].
• Product due diligence: /nike/search (or /nike/suggest) to find a product's slug/style_color, then /nike/product for pricing/sizes and /nike/product/reviews to summarize the rating breakdown and what reviewers say before recommending it.
• Nearest store lookup: /nike/stores with a location's lat/lng to list nearby stores by distance, with address and phone for the closest match.

Notes & limits

• Credits / pay-on-success: billed only on 2xx; free tier 2,000 credits/mo. Key at https://crawlora.net.
• Public data only — public Nike product/category/store pages; respect Nike's terms.
• Security: key lives in CRAWLORA_API_KEY only — never hardcode, query-param, or commit it.
• /nike/search takes exactly one of keyword or category — passing both, or neither, is invalid.
• Product identity is slug + style_color together, not a single SKU or product id — /nike/product and /nike/product/reviews both require the pair, sourced from a search result's colors[].slug / colors[].style_color.
• Keyword search is best-effort relevance, not a guaranteed match: for an obscure or nonsense keyword, Nike's own index falls back to its recommended results instead of an empty list, and the response has no reliable signal to tell a true match from that fallback — don't treat a non-empty keyword result as proof the term was actually found. A punctuation-only/structurally-empty query does return a genuine empty result.
• Results are paginated — pass page to walk /nike/search and /nike/product/reviews; requesting a page past the last one returns a not-found error on both.