Crawlora Ruby SDK

Ruby client for the public Crawlora web-scraping API. It wraps every public endpoint with grouped helpers and a dynamic call interface, plus retries, pagination, middleware hooks, and client-side rate limiting.

Base URL: https://api.crawlora.net/api/v1
Auth: API key (x-api-key) or JWT (Authorization)
Ruby: 3.0+
Operation reference: docs/operations.md · recipes: docs/recipes.md

Install

Published on RubyGems. The current release is a prerelease (1.18.0.pre.sdk.1), so install it with --pre or pin the version:

# Gemfile
gem "crawlora", "1.18.0.pre.sdk.1"

gem install crawlora --pre

Quick start

require "crawlora"

# Reads CRAWLORA_API_KEY from the environment if api_key: is omitted.
client = Crawlora.client(api_key: ENV["CRAWLORA_API_KEY"])

result = client.bing.search(q: "web scraping")
result["data"].each { |item| puts item["title"] }

client.close # release pooled keep-alive connections

Pass a block to auto-close:

Crawlora.client do |client|
  puts client.amazon.product(asin: "B07FZ8S74R", language: "en_US")
end

Calling operations

Grouped helpers map directly to the API (client.<group>.<method>):

client.youtube.video(id: "dQw4w9WgXcQ")
client.google.search(q: "crawlora", country: "US")

Or call any operation dynamically by its id — handy for metaprogramming:

client.request("bing-search", { q: "web scraping", page: 2 })

# Discover operations:
Crawlora::OPERATION_COUNT            # => total operations
Crawlora::GROUPS["bing"]             # => { "search" => "bing-search", ... }
Crawlora::OperationId::BING_SEARCH   # => "bing-search"

Configuration

Crawlora.client(
  api_key: "…",
  timeout: 30,             # seconds per request
  retries: 2,              # retry attempts on retryable failures
  retry_delay: 0.25,       # base backoff (exponential + jitter, honors Retry-After)
  request_id: true,        # attach an x-request-id to every call
  idempotency_keys: true,  # stable Idempotency-Key on POST/PATCH
  rate_limit: 5,           # max requests/second (client-side)
  max_concurrency: 4,      # max in-flight requests across threads
  headers: { "X-Tenant" => "acme" }
)

Per-request overrides use the reserved _-prefixed keyword on grouped calls, or keyword args on request:

client.bing.search(q: "x", _timeout: 5, _headers: { "X-Trace" => "1" })
client.request("bing-search", { q: "x" }, response_type: "text", retries: 0)

Pagination

# Numeric (page/offset) — stops on the first empty page:
client.paginate_items("airbnb-room-reviews", { id: "123" }, max_pages: 5).each do |review|
  puts review["text"]
end

# Cursor mode — supply the cursor param and a next-cursor extractor:
client.paginate("producthunt-leaderboard", {},
                cursor_param: "cursor", next_cursor: ->(page) { page["next_cursor"] }) do |page|
  puts page["data"]
end

Error handling

begin
  client.bing.search(q: "x")
rescue Crawlora::ClientError => e   # 4xx
  warn "rejected (#{e.status}): #{e.message} #{e.code}"
rescue Crawlora::ServerError => e   # 5xx
  warn "server error: #{e.status}"
rescue Crawlora::NetworkError => e  # timeout / transport failure
  warn "network: #{e.message}"
end

All inherit from Crawlora::Error, which exposes status, code, body, raw_body, headers, and request_id.

License

MIT. See LICENSE.