docs: add "Why opencli?" section and comparison guide

[Astro-Han]

Description

Add positioning content to help new visitors understand where opencli fits in the browser automation ecosystem.

Closes #238

Type of Change

• 🐛 Bug fix
• ✨ New feature
• 🌐 New site adapter
• 📝 Documentation
• ♻️ Refactor
• 🔧 CI / build / tooling

Changes

README "Why opencli?" section (both EN and zh-CN)

Added between Highlights and Prerequisites with:

• User-need comparison table (5 scenarios × best tool recommendation)
• 4 core differentiators: zero LLM cost, deterministic, 20+ Chinese platforms, 8 desktop apps
• Link to detailed comparison guide

docs/comparison.md — Comparison Guide

Honest 5-scenario evaluation across 6 tools (Browser-Use, Crawl4AI, Firecrawl, agent-browser, Stagehand, Skyvern):

Scenario	opencli fit
Scheduled batch data extraction	Best
AI agent site operations	Best
Authenticated operations	Best
General web browsing & exploration	Poor (honest)
Desktop app control	Best (unique)

Also includes:

• Key trade-offs section acknowledging adapter maintenance burden
• Complementary usage flow diagram
• Links to developer docs for writing adapters

VitePress sidebar

Added "Comparison" entry under Guide section.

Checklist

• I ran the checks relevant to this PR
• I updated tests or docs if needed
• I included output or screenshots when useful

Documentation (if adding/modifying an adapter)

• Added doc page under docs/ (new docs/comparison.md)
• Updated docs/adapters/index.md table (N/A)
• Updated sidebar in docs/.vitepress/config.mts

Screenshots / Output

N/A — documentation-only changes.

[Astro-Han]

Preview

README "Why opencli?" section

There are many great browser automation tools. Here's when opencli is the right choice:

Your need	Best tool	Why
Scheduled data extraction from specific sites	opencli	Pre-built adapters, deterministic JSON, zero LLM cost
AI agent needs reliable site operations	opencli	300+ commands, structured output, sub-second response
Explore an unknown website ad-hoc	Browser-Use, Stagehand	LLM-driven general browsing for one-off tasks
Large-scale web crawling	Crawl4AI, Scrapy	Purpose-built for throughput and scale
Control desktop Electron apps from terminal	opencli	CDP + AppleScript — the only CLI tool that does this

What makes opencli different:

• Zero LLM cost — No tokens consumed at runtime. Run 10,000 times and pay nothing.
• Deterministic — 55 sites, 300+ commands. Same command, same output schema, every time.
• 20+ Chinese platforms — Bilibili, Zhihu, Xiaohongshu, Douban, Weibo, and more — unmatched coverage.
• 8 desktop apps — Cursor, Codex, Notion, ChatGPT, Discord via CDP — unique in the ecosystem.

---

docs/comparison.md — Scenario 4 (honest "Poor" rating)

Tool	Fit	Notes
Browser-Use	Best	LLM-driven, handles arbitrary websites and flows.
Stagehand	Best	Clean API for act(), extract(), observe() on any page.
agent-browser	Good	Token-efficient primitives for AI agents.
Skyvern	Good	Visual AI that generalizes across sites.
opencli	Poor	Only works with sites that have pre-built adapters. Cannot handle arbitrary websites.

---

docs/comparison.md — Complementary usage flow

Has adapter?  ──yes──▶  opencli (fast, free, deterministic)
     │
     no
     │
     ▼
One-off task?  ──yes──▶  Browser-Use / Stagehand (LLM-driven)
     │
     no
     │
     ▼
Recurring?    ──yes──▶  Write an opencli adapter, then use opencli