Search King

An all-in-one human-machine search tool—image search, text search, web reading, deep crawling, and anti-bot site scraping. One machine, eighteen engines.

Five Behaviors

Image / text / read / crawl / platform fetch

Engine Matrix

12 image engines + 6 text engines

Six Backends

From headless browser to real browser, layered fallbacks

Workflow

Switch on CAPTCHA; fall back to real-human mode if all else fails

Traceable Copyright

Every result carries its source and license record

Quick Start

One command, install and go

One Machine, Five Ways of Working

Behavior parameters combine orthogonally—one tool covers every content-collection scenario.

Behavior, backend, and decorator parameters combine freely without interfering—which also keeps the command line predictable and automatable.

  • Image search (default): type a keyword and search right away, with filters by engine and by free sites.
  • Text search: `--search`, runs the 6-engine text chain, and can skip the browser for a fast pure-HTTP search.
  • Read web: `--read`, extracts the main text, with parallel support for multiple URLs.
  • Deep crawl: `--crawl`, deep extraction based on Crawl4AI, output as Markdown.
  • Platform fetch: `--fetch`, targeted at Xueqiu posts and Bilibili videos and collections.

Eighteen Engines, the Best One Wins

Twelve image + six text, covering mainstream and regional search engines.

  • Image (12): Unsplash, Pexels, Pixabay, Burst, Bing, Baidu, Brave, DuckDuckGo, Yandex, Sogou, 360, Google.
  • Text (6): Bing, Baidu, Google, DuckDuckGo, Brave, Sogou.
  • Free-for-commercial first: `--free` only uses photography sites with a free commercial license, keeping copyright risk under control.
  • Automatic engine-chain fallback: if one engine goes down or gets blocked, it moves to the next automatically without stopping.

Six Backends, Layered Fallbacks

From lightest to heaviest, escalating by how blocked the site is.

  • edge: the system Edge browser (default).
  • chromium: Playwright's built-in Chromium, consistent across platforms.
  • cloak: CloakBrowser anti-detection, built to get past Cloudflare-class protection.
  • lite: pure HTTP, the fastest, only for text search.
  • opencli: real Edge + persistent session, resists WAF, for Xueqiu and Bilibili.
  • manual: you browse, the tool extracts—when anti-bot fights back, a real human is the final fallback.

Blocked? Take Another Route

Skip CAPTCHAs automatically, fall back engine by engine, and let a real human finish it.

Search King treats "being blocked" as the norm: any mode that hits a CAPTCHA or a block skips to the next engine automatically; if all of them fail, it switches to manual mode for a human to finish.

  • Auto-skip CAPTCHAs: never freeze or submit—just switch routes.
  • Deep-crawl to documents: crawl results land as Markdown, ready for further processing.
  • Concurrent downloads: batch-download selected images, always carrying source and license records.

Copyright, Not an Afterthought

The age of solving copyright with one stock image is over—sources must be checkable, licenses traceable.

Downloading an image isn't the end of the job; it's delivering a usable asset. Search King always carries source documents and license info, so every download can say where the image came from and whether it can be used commercially.

Install, and Go

Install the dependencies, and one command gets you running.

python scraper.py "cat"
python scraper.py "cat" --free --download -o pics
python scraper.py --search "AI 2026" --backend lite
python scraper.py --read https://example.com --backend cloak
Playwright Crawl4AI CloakBrowser yt-dlp Python 3.8+
View the GitHub repo Back to Tech Work

Wants to build it, and can—technology isn't the end, it's turning a nuisance into a tool

Back to Tech Work →