Scraperapi Ruby Sdk

ScraperAPI — Ruby SDK Best Practices

Requires: Ruby >= 2.0, gem install scraperapi (or gem 'scraperapi' in Gemfile), SCRAPERAPI_API_KEY environment variable.

Setup

require "scraper_api"

client = ScraperAPI::Client.new(ENV["SCRAPERAPI_API_KEY"])

Never hardcode the API key. Read it from the environment every time.

Basic Usage

# Simple GET — returns raw HTML string via .raw_body
html = client.get("https://example.com/").raw_body
puts html

# With a single parameter
html = client.get("https://example.com/", render: true).raw_body

# With multiple parameters
html = client.get(
  "https://example.com/",
  render: true,
  country_code: "us"
).raw_body

Parameters are passed as keyword arguments after the URL. .raw_body extracts the HTML string from the response object.

Decision Guide

| Situation | Approach | |-----------|---------| | Single URL, synchronous | client.get(url, **params).raw_body | | Page loads content via JavaScript | Pass render: true | | Site blocks datacenter proxies | Pass premium: true | | Toughest anti-bot protection | Pass ultra_premium: true | | Multi-step / paginated flow on same domain | Use session_number: | | 20+ URLs or batch jobs | Use async endpoint via Net::HTTP or Faraday | | Supported platform (Amazon, Google, etc.) | Use structured data endpoint directly |

Parameter Reference

Rendering

# Render JavaScript before returning HTML
# Use when: page is a React/Vue/Angular SPA, or initial scrape returns empty/partial content
# Cost: +10 credits
html = client.get("https://spa-site.com/", render: true).raw_body

# Wait for a specific DOM element (requires render: true)
html = client.get(
  "https://spa-site.com/",
  render: true,
  wait_for_selector: ".product-list"
).raw_body

Don't add render: true by default — try without it first. It increases cost and latency.

Proxies and Geotargeting

# Route through a country-specific proxy — no extra credit cost
html = client.get("https://example.com/", country_code: "gb").raw_body

# Premium residential/mobile IPs — for sites that block datacenter proxies
# Cost: 10 credits (25 with render: true)
html = client.get("https://hard-site.com/", premium: true).raw_body

# Ultra-premium — for the toughest anti-bot protections
# Cost: 30 credits (75 with render: true)
# Note: incompatible with keep_headers — custom headers are discarded
html = client.get("https://hardest-site.com/", ultra_premium: true).raw_body

premium and ultra_premium are mutually exclusive — never pass both. Escalation order: standard (1 cr) → render (10 cr) → premium (10 cr) → ultra_premium (30 cr).

Sessions (Sticky Proxy)

# Reuse the same proxy IP across requests — useful for pagination and multi-step flows
# Sessions expire 15 minutes after last use; any integer is a valid session ID
html1 = client.get("https://example.com/page1", session_number: 42).raw_body
html2 = client.get("https://example.com/page2", session_number: 42).raw_body

Headers and Device Type

# Forward custom headers to the target site
# Note: keep_headers is ignored when ultra_premium: true
html = client.get(
  "https://example.com/",
  keep_headers: true
  # Pass additional headers via the underlying request object as needed
).raw_body

# Emulate a mobile or desktop browser user-agent
html = client.get("https://example.com/", device_type: "mobile").raw_body

Autoparse

# Return structured JSON instead of HTML for supported sites
# Use for Amazon, Google, and other supported platforms when you want clean data
json_result = client.get("https://amazon.com/dp/B09V3KXJPB", autoparse: true).raw_body

Escalation Ladder

Always start with the cheapest option and escalate only when blocked.

def scrape_with_escalation(client, url)
  tiers = [
    {},
    { render: true },
    { premium: true },
    { premium: true, render: true },
    { ultra_premium: true },
  ]

  tiers.each do |params|
    result = client.get(url, **params).raw_body
    return result if result&.include?(" "application/json")
  JSON.parse(response.body) # { "id" => "...", "statusUrl" => "..." }
end

def poll_job(job, max_wait: 120, interval: 5)
  deadline = Time.now + max_wait
  while Time.now  e
  status = e.respond_to?(:response) ? e.response&.code&.to_i : nil
  case status
  when 401 then raise "Invalid API key — check SCRAPERAPI_API_KEY"
  when 403 then raise "Blocked or out of credits — try premium: true or ultra_premium: true"
  when 429 then raise "Rate limit hit — reduce concurrency or switch to async"
  when 500, 503 then raise "Transient error — retry with exponential backoff"
  else raise
  end
end

Status code reference: 200 success, 401 bad key, 403 blocked/no credits, 404 target not found, 429 rate limit, 500/503 transient (not charged — safe to retry).

Credit Cost Reference

| Request type | Credits | |---|---| | Standard | 1 | | render: true | 10 | | premium: true | 10 | | premium: true, render: true | 25 | | ultra_premium: true | 30 | | ultra_premium: true, render: true | 75 |

Documentation

• Ruby SDK getting started
• SDK method reference
• JavaScript rendering
• Sessions
• API status codes
• Dashboard & credits

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: scraperapi
• Source: scraperapi/scraperapi-skills
• License: MIT

Install and usage instructions live in the source repository linked above.