ShotAPI MCP Server

ShotAPI — Web Screenshot & Render MCP Server for AI Agents

Give your AI Agent eyes. One command, zero install.

中文文档

Quick Start (Remote Mode — No Install)

# Claude Code
claude mcp add --transport streamable-http shotapi https://aiphotoshop.mynatapp.cc/mcp

# Or with npx
npx -y @anthropic-ai/claude-code@latest --mcp https://aiphotoshop.mynatapp.cc/mcp

For Claude Desktop / Cursor, add to your config:

{
  "mcpServers": {
    "shotapi": {
      "type": "streamable-http",
      "url": "https://aiphotoshop.mynatapp.cc/mcp"
    }
  }
}

Real Use Cases

1. Claude verifies its own code

Human: Build a landing page with a hero section, testimonials, and pricing grid.

Claude: [writes HTML/CSS code]

Claude: Let me check this with the render tool...

[Claude calls ShotAPI render tool with the HTML it wrote]

Claude: The pricing grid overlaps on mobile. Fixing responsive breakpoints...

2. Deployment verification

Human: I just deployed. Check if it looks right.

Claude: [calls ShotAPI screenshot on https://yourapp.com]

Claude: Footer text is cut off — CSS overflow issue. Background image 404.

3. UI regression in CI/CD

# Capture before/after screenshots and compare
curl -s "https://aiphotoshop.mynatapp.cc/v1/screenshot?url=https://staging.myapp.com" -o before.png
curl -s "https://aiphotoshop.mynatapp.cc/v1/screenshot?url=https://staging.myapp.com" -o after.png

3 MCP Tools

Tool	Description	Example
screenshot_one_liner	One URL, one screenshot	screenshot_one_liner(url="https://github.com")
screenshot	Full control: viewport, full-page, CSS selector, format	screenshot(url="https://github.com", selector="#hero")
render	Render HTML/CSS to image — Agent writes code, sees result instantly	render(html="<h1>Hello</h1>")

The render tool closes the feedback loop: write code → render → see → revise → render again.

Why ShotAPI?

• MCP-native — No glue code. Agent calls tools directly in conversation.
• Zero install remote mode — streamable-http, no Python/Playwright needed.
• Direct China access — No VPN required.
• Free tier — 50 screenshots + 50 renders/month, IP-based, no signup, no API key. All features unlocked (full-page, selector, WebP).
• Render capability — Not just screenshots of existing pages. Render Agent-generated HTML.
• Built-in ad blocking — Cleaner screenshots, less noise.
• Listed in MCP Registry — Official Anthropic MCP server directory.

Try It Now (No Signup)

curl -s "https://aiphotoshop.mynatapp.cc/v1/screenshot?url=https://github.com" -o shot.jpg

Pricing

Plan	Price	Limit
Free	$0	50+50/month (IP-based, all features, no signup)
Starter	$4.90/mo	5,000/month
Pro	$9.90/mo	20,000/month

Free tier: no API key needed. Paid: get key at pricing page.

Local Mode (STDIO)

For private deployments or paid-tier usage, run locally:

# Install dependencies
pip install -r requirements.txt
playwright install chromium

# Set env vars
export SHOTAPI_BASE_URL=https://aiphotoshop.mynatapp.cc
export SHOTAPI_KEY=your_key_here  # optional for paid tiers

# Add to Claude Code
claude mcp add shotapi python mcp_stdio.py

Links

• Docs: https://aiphotoshop.mynatapp.cc/docs
• Pricing: https://aiphotoshop.mynatapp.cc/pricing
• Health: https://aiphotoshop.mynatapp.cc/health
• MCP Registry: https://registry.modelcontextprotocol.io/servers/io.github.smallhandsome/shotapi-mcp-server
• Smithery: https://smithery.ai/server/@ljs/shotapi
• awesome-mcp-servers: https://github.com/punkpeye/awesome-mcp-servers
• Dev.to Article: https://dev.to/smallhandsome/give-your-ai-agent-eyes-building-a-visual-mcp-server-for-web-screenshots-3f49