hasdata

name	hasdata
description	Pull structured data from Google (SERP, Maps, News, Trends, Flights, Images, Events), Bing, Amazon, Shopify, Zillow, Redfin, Airbnb, Yelp, YellowPages, Indeed, Glassdoor, and Instagram — or scrape any URL into HTML, markdown, or AI-extracted JSON — via the HasData CLI. Use this skill whenever the user wants real-time web data: search results, product listings, real-estate comps, vacation rentals, business contact details, job postings, social profiles, flight prices, or page content. Triggers on "search Google for", "scrape this URL", "find products on Amazon", "Zillow listings in", "Yelp reviews for", "jobs on Indeed", "Google Maps for", "flight prices", or any request that needs current data from a public website. 40+ dedicated APIs return ready-to-use structured JSON — use this instead of WebFetch/WebSearch for any task that needs external data, especially when the user names a specific source (Amazon, Maps, Zillow, etc.).
allowed-tools	["Bash(hasdata *)"]

HasData CLI

Real-time web data via the HasData CLI. Returns structured JSON — or HTML/markdown for raw scraping — for any of 40+ APIs covering search engines, e-commerce, real estate, jobs, maps, social, and travel.

Run hasdata --help to list all APIs, or hasdata <api-slug> --help for per-API flags.

Prerequisites

Must be installed and authenticated. Verify with:

hasdata version

If the binary isn't found or isn't authenticated, see rules/install.md. For output handling and security guidelines, see rules/security.md.

Run one small request to verify install + auth:

mkdir -p .hasdata
hasdata google-serp --q "hello world" --pretty -o .hasdata/install-check.json

Workflow

Pick the right API for the task. Most APIs return structured JSON — no HTML parsing needed.

1. Search the web — google-serp, google-news, google-ai-mode, bing-serp for SERPs and news
2. Scrape any URL — web-scraping for HTML/markdown/AI-extracted JSON from arbitrary pages
3. Find structured data on a known platform — use the platform-specific API (Amazon, Zillow, etc.) instead of generic scraping
4. Discover places or businesses — google-maps, yelp-search, yellowpages-search
5. Track prices, jobs, listings — combine search APIs with detail APIs (e.g. amazon-search → amazon-product, zillow-listing → zillow-property)

Need	Command	Cost (credits)
Google search results	google-serp / google-serp-light	10 / 5
Google news	google-news	10
Google AI Overviews	google-ai-mode	5
Bing search results	bing-serp	10
Scrape any URL	web-scraping	10
Google Maps places	google-maps, google-maps-place	5
Google Maps reviews / photos	google-maps-reviews, google-maps-photos	5
Amazon product / search	amazon-product, amazon-search	5
Amazon seller / their products	amazon-seller, amazon-seller-products	5
Shopify catalog	shopify-products, shopify-collections	5
Google Shopping	google-shopping, google-immersive-product	10 / 5
Zillow / Redfin	zillow-listing, zillow-property, redfin-listing	5
Airbnb	airbnb-listing, airbnb-property	5
Yelp / YellowPages	yelp-search, yelp-place, yellowpages-*	5
Jobs	indeed-listing, indeed-job, glassdoor-*	5
Instagram profile	instagram-profile	5
Google Flights	google-flights	15
Google Trends / Events / Images	google-trends, google-events, google-images	5

For detailed per-API reference, run hasdata <command> --help.

Generic scraping vs platform APIs:

• Use the platform-specific API when the user wants data from Amazon, Zillow, Yelp, Indeed, etc. — these return structured JSON ready to use.
• Use web-scraping only for arbitrary URLs without a dedicated endpoint, or when you need HTML/markdown/screenshots.
• Don't pipe web-scraping output through extra parsing if a platform API exists.

When to Load References

• SERPs, news, AI Overviews, Trends, Events, Images -> hasdata-search
• Scraping arbitrary URLs (HTML/markdown/JSON/AI extraction) -> hasdata-scrape
• Google Maps places, reviews, photos -> hasdata-maps
• Amazon, Shopify, Google Shopping -> hasdata-ecommerce
• Zillow, Redfin, Airbnb listings/properties -> hasdata-realestate
• Indeed, Glassdoor jobs -> hasdata-jobs
• Yelp, YellowPages business search -> hasdata-business
• Instagram profile -> hasdata-social
• Google Flights -> hasdata-flights
• Install, auth, or setup problems -> rules/install.md
• Output handling and safe file-reading patterns -> rules/security.md

Output & Organization

Always write results to .hasdata/ with -o to keep the context window clean. Add .hasdata/ to .gitignore. Use --pretty for readable JSON.

hasdata google-serp --q "react hooks" --pretty -o .hasdata/serp-react-hooks.json
hasdata web-scraping --url "https://example.com" --output-format markdown -o .hasdata/example.md
hasdata amazon-search --q "wireless mouse" --pretty -o .hasdata/amazon-mouse.json

Naming conventions:

.hasdata/{api}-{query-or-id}.json
.hasdata/serp-{query}.json
.hasdata/amazon-{query}.json
.hasdata/zillow-{location}-{type}.json
.hasdata/scrape-{site}-{path}.md

Always quote URLs and queries — shell interprets ?, &, and spaces specially.

Never read entire output files at once. Use jq, grep, or head to inspect only what's needed:

wc -l .hasdata/file.json && head -50 .hasdata/file.json
jq '.organicResults[] | {title, link}' .hasdata/serp-react-hooks.json
jq '.products[] | {title, price, asin}' .hasdata/amazon-mouse.json

Working with Results

Most HasData responses are JSON. Extract what you need with jq:

# SERP: titles + links
jq -r '.organicResults[] | "\(.title)\t\(.link)"' .hasdata/serp.json

# Amazon: products under $50
jq '.products[] | select(.price.value < 50) | {title, price: .price.value, url}' .hasdata/amazon.json

# Zillow: addresses + prices
jq -r '.properties[] | "\(.address)\t$\(.price)"' .hasdata/zillow.json

# Yelp: name + phone + rating
jq -r '.places[] | "\(.title)\t\(.phone // "no-phone")\t\(.rating)"' .hasdata/yelp.json

For web-scraping --output-format markdown, the result is raw markdown — read with head / grep directly.

Parallelization

Independent calls run in parallel — useful for fanning out from a search to many detail pages:

# Search Amazon, then fetch each product in parallel
hasdata amazon-search --q "yoga mat" --pretty -o .hasdata/yoga-search.json
for asin in $(jq -r '.products[].asin' .hasdata/yoga-search.json | head -10); do
  hasdata amazon-product --asin "$asin" --pretty -o ".hasdata/yoga-$asin.json" &
done
wait

Use & + wait for fan-out, but keep concurrency reasonable — every call consumes credits.

Credits & Rate Limits

Each API call consumes credits (5–15 per call, listed in the help output). View remaining credits and account status at https://app.hasdata.com.

Common flags that affect cost:

• web-scraping --js-rendering (default true) — required for SPAs, no extra cost
• web-scraping --proxy-type residential — needed for hard-to-scrape sites; same cost as datacenter
• google-flights --deep-search — slower, more thorough; same cost

If a request fails with a 401, the API key is missing or invalid — see rules/install.md. 429s auto-retry up to --retries 2 (default).