LLM Scraper MCP

Natural language web data extraction. No CSS selectors, no coding.

Describe what data you want in plain language — the AI understands the page and extracts structured JSON for you. No CSS selectors, no XPath, no coding. Bring your own LLM API key (DeepSeek / OpenAI / Kimi supported).

Features

• extract_data: natural language → structured JSON
• extract_with_schema: define the exact output structure
• extract_links: all links with anchor text
• export_csv: convert results to CSV
• Auto-handles all levels of anti-bot protection
• Works with DeepSeek / OpenAI / Kimi API keys

Quick Start

1. 1. pip install mcp patchright
2. 2. Add to MCP config with env:
3. LLM_API_KEY=sk-xxx, LLM_API_BASE=https://api.deepseek.com/v1/chat/completions, LLM_MODEL=deepseek-chat
4. 3. Ask the AI:
5. "extract all product names and prices from https://example.com"

Best for: AI agent users · Analysts · No-code data extraction