BENCHMARK // ARCHITECTURE EVALUATION

@jackwener/opencli vs. Scrapfly

A rigorous comparative analysis of local browser bridge automation vs. cloud anti-bot scraping APIs for Google Antigravity autonomous workflows.

@jackwener/opencli Local Runtime

"Make Any Website & Tool Your CLI." Connects Antigravity to active logged-in Chrome profiles and Electron desktop apps with zero API fees.

  • Execution: Local Node.js daemon (:19825) + Chrome Bridge extension.
  • Auth Model: 100% reuses authentic logged-in user profile, cookies, and local IP.
  • Action Range: Full read & write (posting, form fills, clicks, media downloads).
  • Cost / License: 100% Free & Open-Source (Apache-2.0).
Scrapfly.io Cloud API

"Scraping Without Getting Blocked." Managed cloud scraping and anti-bot bypass API with global residential proxies and LLM Markdown preprocessing.

  • Execution: Managed serverless REST API + headless cloud container fleet.
  • Anti-Bot: Automated ASP (Cloudflare Turnstile, DataDome, Akamai, Kasada bypass).
  • Features: 130M+ residential proxies, C++ Curlium engine, LLM Markdown extraction.
  • Cost / License: Commercial metered API credits ($30 to $500+/mo).

📊 Detailed Head-to-Head Comparison Matrix

Dimension @jackwener/opencli Scrapfly
Primary Target Personal AI agent automation, logged-in session manipulation, desktop apps. High-volume public data extraction, hostile anti-bot bypass, batch ETL.
Execution Environment Local Machine: Node.js daemon (:19825) + Chrome Extension / CDP. Managed Cloud: Serverless REST API endpoints.
Anti-Bot Strategy Authentic User Session: Native browser fingerprint, local IP, valid cookies (Zero alarms). Adversarial ASP: TLS/JA4 spoofing, automated CAPTCHA solver, 130M+ residential proxies.
Action Spectrum Read & Write: Click, fill, type, post, like, download, inspect state. Read-Only: DOM extraction, JS rendering, screenshot capture, LLM Markdown.
Authentication Reuses active Chrome profile (--profile work); zero credentials stored. Stateless API keys; custom cookie/header injection required for private pages.
Token Optimization Outputs compact structured data (-f json) + pruned ref-tree DOM states. Server-side format=markdown strips 80%+ of HTML token boilerplate.
Cost & Licensing 100% Free (Apache-2.0 open-source, $0 compute/API cost). Metered SaaS ($30–$500+/mo based on credits consumed).
Concurrency Scale Vertical single-machine scale (5–20 tabs before local RAM/CPU limits). Horizontal cloud scale (5 to 100+ concurrent requests across global nodes).
Data Privacy 100% Zero-Knowledge: Cookies, tokens, and payloads never leave localhost. Third-Party Cloud: URLs and payloads route through Scrapfly nodes.

🎯 Antigravity Suitability & Decision Framework

For Google Antigravity (AGY / Gemini core), we recommend a Unified Hybrid Architecture:

When to Use @jackwener/opencli

  • Authenticated Social / SaaS Actions: Interacting with logged-in services (Twitter/X, Xiaohongshu, Bilibili, Reddit, Jira, Notion) without needing developer API keys.
  • Local Dev & Intranets: Automating http://localhost:3000 or corporate VPN endpoints.
  • Desktop Electron App Automation: Driving Cursor, Antigravity IDE, or Discord via CDP port 9222.
  • Zero-Cost Daily CLI Queries: Standardized extraction for daily routines with zero API bills.

When to Use Scrapfly

  • Hostile Anti-Bot Targets: Scraping domains protected by Cloudflare Turnstile, Under Attack Mode, DataDome, or Akamai.
  • Mass Parallel Research (parallel-search): Offloading 10–30 simultaneous URL fetches to cloud workers without local CPU/RAM load.
  • Headless / Remote Cloud Environments: Running Antigravity inside CI/CD, Docker containers, or remote Linux servers.
  • Clean Markdown Preprocessing: Trimming 80%+ of HTML token overhead server-side.