webserp 0.1.3

Metasearch CLI — query multiple search engines in parallel with browser impersonation

webserp

Metasearch CLI — query multiple search engines in parallel with browser impersonation.

Like grep for the web. Searches Google, Bing, DuckDuckGo, Brave, Yahoo, Mojeek, Startpage, and Presearch simultaneously, deduplicates results, and returns clean JSON.

Why webserp?

Most search scraping tools get rate-limited and blocked because they use standard HTTP libraries. webserp uses curl_cffi to impersonate real browsers (Chrome TLS/JA3 fingerprints), making requests indistinguishable from a human browsing.

• 8 search engines queried in parallel
• Browser impersonation via curl_cffi — bypasses bot detection
• Fault tolerant — if one engine fails, others still return results
• SearXNG-compatible JSON output format
• No API keys — scrapes search engine HTML directly
• Fast — parallel async requests, typically completes in 2-5s

Install

pip install webserp

Usage

# Search all engines
webserp "how to deploy docker containers"

# Search specific engines
webserp "python async tutorial" --engines google,brave,bing

# Limit results per engine
webserp "rust vs go" --max-results 5

# Show which engines succeeded/failed
webserp "test query" --verbose

# Use a proxy
webserp "query" --proxy "socks5://127.0.0.1:1080"

Output Format

JSON output matching SearXNG's format:

{
  "query": "deployment issue",
  "number_of_results": 42,
  "results": [
    {
      "title": "How to fix Docker deployment issues",
      "url": "https://example.com/docker-fix",
      "content": "Common Docker deployment problems and solutions...",
      "engine": "google"
    }
  ],
  "suggestions": [],
  "unresponsive_engines": []
}

Options

Flag	Description	Default
-e, --engines	Comma-separated engine list	all
-n, --max-results	Max results per engine	10
--timeout	Per-engine timeout (seconds)	10
--proxy	Proxy URL for all requests	none
--verbose	Show engine status in stderr	false
--version	Print version

Engines

google, bing, duckduckgo, brave, yahoo, mojeek, startpage, presearch

For OpenClaw and AI agents

Built for AI agents. Tools like OpenClaw and other AI agents need reliable web search without API keys or rate limits. webserp uses curl_cffi to mimic real browser fingerprints — results like a browser, speed like an API. It queries 8 engines in parallel, so even if one gets rate-limited, results still come back.

Why a CLI tool instead of a Python library?

A CLI tool keeps web search out of the agent's process. The agent calls webserp, gets JSON back, and the process exits — no persistent HTTP sessions, no in-process state, no import overhead. Agents that never need web search pay zero cost.

Example agent use cases

• Research — searching the web for current information before answering user questions
• Fact checking — verifying claims against multiple search engines
• Link discovery — finding relevant URLs, documentation, or source code
• News monitoring — checking for recent events or updates on a topic

# Agent searching for current information
webserp "latest python 3.14 release date" --max-results 5

# Searching multiple engines for diverse results
webserp "docker networking troubleshooting" --engines google,brave,bing --max-results 3

# Quick search with verbose to see which engines responded
webserp "CVE-2024 critical vulnerabilities" --verbose --max-results 5

License

MIT