⚡ OFFICIAL CONNECTORS & MODEL CONTEXT PROTOCOL (MCP)

Connect FlyCrawl with Any
AI Framework, Agent or IDE

Feed high-speed, noise-free Web Markdown directly into your RAG pipelines and autonomous agents. Native MCP servers, zero-latency vector generation, and plug-and-play recipes.

8+ Verified Connectors
100% MCP Protocol Ready
-92% Token Noise Slashed
<350ms Average Latency 
# 1. Install battle-tested standard SDK (Zero Vendor Lock-In)
pip install firecrawl-py

# 2. Extract clean Fit-Markdown via FlyCrawl high-speed engine
from firecrawl import FirecrawlApp

app = FirecrawlApp(api_key="fly_live_YOUR_API_KEY", api_url="https://api.flycrawl.net/v1")
doc = app.scrape_url("https://en.wikipedia.org/wiki/Artificial_intelligence")
print(doc["markdown"][:300])
// 1. Install standard client (Zero Vendor Lock-In)
npm install @mendable/firecrawl-js

// 2. Connect TypeScript app to FlyCrawl wire-compatible API
import FirecrawlApp from '@mendable/firecrawl-js';

const app = new FirecrawlApp({ apiKey: 'fly_live_YOUR_API_KEY', apiUrl: 'https://api.flycrawl.net/v1' });
const result = await app.scrapeUrl('https://example.com');
console.log(result.markdown);
// Add directly to ~/.cursor/mcp.json or Claude Desktop config
{
  "mcpServers": {
    "flycrawl": {
      "command": "npx",
      "args": ["-y", "flycrawl-mcp"],
      "env": {
        "FLYCRAWL_API_KEY": "fly_live_YOUR_API_KEY"
      }
    }
  }
}
# Native REST API with sub-350ms response
curl -X POST https://api.flycrawl.net/v1/scrape \
  -H "Authorization: Bearer fly_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com", "formats": ["markdown", "chunks"]}'
🔍
Showing all 8 verified integrations
OFFICIAL SDK

LangChain

Load live websites, multi-page crawls, and sitemaps directly as LangChain Document objects. Slashes 92% of HTML bloat and scripts before chunking.

Python TypeScript RAG Loader <350ms
python 
from langchain_community.document_loaders import FlyCrawlLoader
loader = FlyCrawlLoader(api_key="fly_live_...", url="https://site.com")
docs = loader.load()
View LangChain Recipe
OFFICIAL READER

LlamaIndex

Ingest clean, fit-to-context web knowledge into LlamaIndex VectorStoreIndex. Build ultra-accurate Q&A query engines with zero hallucination.

Python VectorStoreIndex Hybrid RAG
python 
from llama_index.readers.flycrawl import FlyCrawlReader
reader = FlyCrawlReader(api_key="fly_live_...")
documents = reader.load_data(url="https://docs.target.com")
View LlamaIndex Recipe
MCP PROTOCOL

Cursor IDE

Enable Cursor Composer and Agent mode to browse live documentation, inspect GitHub code, and crawl APIs autonomously via native MCP server.

MCP Server Agent Mode Windsurf Safe
mcp.json 
{
  "command": "npx",
  "args": ["-y", "flycrawl-mcp"],
  "env": { "FLYCRAWL_API_KEY": "fly_live_..." }
}
View Cursor Setup Guide
MCP DESKTOP

Claude Desktop

Equip Claude 3.7 Sonnet with native web browsing in Claude Desktop. Ask Claude to crawl and summarize live web pages instantly.

Claude 3.7 Desktop Tool Model Context
claude_desktop_config.json 
"flycrawl": {
  "command": "npx",
  "args": ["-y", "flycrawl-mcp"],
  "env": { "FLYCRAWL_API_KEY": "fly_live_..." }
}
View Claude Guide
OFFICIAL COMMUNITY NODE

n8n (Official Node)

Official n8n-nodes-flycrawl community node. Scrape dynamic web pages, generate 1536/3072-dim vectors for RAG, and empower autonomous AI Agents in n8n.

Community Node npm Package 1536-dim Vectors AI Agent Tool
n8n Settings > Community Nodes 
# Install in n8n Community Nodes:
npm install n8n-nodes-flycrawl
# Or GUI: Install 'n8n-nodes-flycrawl'
View n8n Guide & Node Setup
AGENT TOOL

Dify.ai

Equip Dify conversational agents and LLM workflows with live web extraction. Plug-and-play OpenAPI YAML tool definition included.

OpenAPI Tool Agent YAML Dify Cloud
dify_tool.yaml 
identity:
  name: flycrawl_scraper
  author: FlyCrawl
description:
  human: Live web scraper for LLM markdown
View Dify YAML Recipe
VISUAL CANVAS

Flowise AI

Drag and drop web scraping capabilities into your visual Flowise canvas. Connect directly with vector stores and autonomous agents.

Custom Tool Visual Flow Zero-Code
Flowise Custom Tool 
const response = await fetch('https://api.flycrawl.net/v1/scrape', {
  method: 'POST',
  headers: { 'Authorization': 'Bearer fly_live_...' },
  body: JSON.stringify({ url: $url })
});
View Flowise Canvas Setup
FUNCTION CALLING

OpenAI Function Calling

Equip GPT-4o and o3-mini models with native web searching and scraping schemas. Returns pure Markdown without third-party proxy lag.

GPT-4o JSON Schema Zero Bloat
tools.json 
{
  "type": "function",
  "function": {
    "name": "flycrawl_scrape",
    "parameters": { "properties": { "url": { "type": "string" } } }
  }
}
View OpenAI Schema
⚡ INSTANT SETUP GENERATOR

Generate Ready-to-Paste MCP Client Configuration

Select your AI coding assistant or client below to get the exact JSON snippet formatted and ready to paste into your configuration file.

~/.cursor/mcp.json
🌟 WHY FLYCRAWL FOR AI?

Engineered Specifically for AI Context Windows

Raw HTML drains 90% of your LLM token budget on boilerplate. FlyCrawl converts raw websites into distilled, token-efficient Fit-Markdown in sub-350ms.

⚡

-92% Token Compression

Strips navigation bars, cookie banners, tracking scripts, and stylesheets. Keeps only semantic headers, tables, code blocks, and core copy.

🛡️

Cloudflare & Anti-Bot Armor

Autonomous cloud browser pool with fingerprint rotation, residential proxies, and TLS stealth handles heavy JavaScript SPAs effortlessly.

🧠

1536 & 3072-dim Embeddings

Optionally generate vector embeddings right in the scraping response for Pinecone, Qdrant, and Milvus with zero latency overhead.

🔌

Native Model Context Protocol

Compliant with Anthropic's Model Context Protocol. Works out of the box with Claude Desktop, Cursor Composer, Windsurf, and Goose.