Sign in Subscribe
SharingExperience

Perbedaan Firecrawl Self-Host v1 vs v2 dan Cara Akses API & MCP

Perbedaan Firecrawl Self-Host v1 vs v2 dan Cara Akses API & MCP

1️⃣ Perbedaan Firecrawl Self-Host v1 vs v2

Perbedaan Utama

Aspek v1 v2
URL API https://api.firecrawl.dev/v1/ https://api.firecrawl.dev/v2/
Caching Tidak default Default 2 hari (maxAge)
Default Options Tidak ada blockAds: true, skipTlsVerification: true, removeBase64Images: true
Summary Format Tidak ada ✅ "summary" tersedia
JSON Extraction "extract" { type: "json", prompt, schema }
Screenshot Format Sederhana { type: "screenshot", fullPage, quality, viewport }
Search Sources Web saja "web", "news", "images"
Crawl Prompt Tidak ada ✅ Natural language prompt dengan /crawl/params-preview

Perubahan Nama Method (SDK)

JavaScript/TypeScript:

v1 (FirecrawlApp) v2 (Firecrawl)
scrapeUrl(url, ...) scrape(url, options?)
search(query, ...) search(query, options?)
mapUrl(url, ...) map(url, options?)
crawlUrl(url, ...) crawl(url, options?)
asyncCrawlUrl(...) startCrawl(url, options?)
checkCrawlStatus(id) getCrawlStatus(id)

Python:

v1 v2
scrape_url(...) scrape(...)
map_url(...) map(...)
crawl_url(...) crawl(...)
async_crawl_url(...) start_crawl(...)
check_crawl_status(...) get_crawl_status(...)

Perubahan Crawl Options

v1 v2
allowBackwardCrawling Gunakan crawlEntireDomain
maxDepth Gunakan maxDiscoveryDepth
ignoreSitemap (bool) `sitemap: "only"
- ✅ prompt (natural language)

2️⃣ Cara Akses API Firecrawl

A. API Firecrawl Cloud

Setup:

  1. Daftar di firecrawl.dev
  2. Dapatkan API key dari dashboard
  3. Gunakan base URL: https://api.firecrawl.dev/v2/

Contoh Penggunaan (Python):

# Install
pip install firecrawl-py

# Kode
from firecrawl import Firecrawl

firecrawl = Firecrawl(api_key="fc-YOUR-API-KEY")

# Scrape single URL
doc = firecrawl.scrape(
    "https://firecrawl.dev",
    formats=["markdown", "html"],
)

# Crawl website
response = firecrawl.crawl(
    "https://firecrawl.dev",
    limit=100,
    scrape_options={"formats": ["markdown", "html"]},
)

Contoh Penggunaan (JavaScript):

import Firecrawl from '@mendable/firecrawl-js';

const firecrawl = new Firecrawl({ apiKey: 'fc-YOUR-API-KEY' });

const doc = await firecrawl.scrape('https://firecrawl.dev', {
  formats: ['markdown', 'html'],
});

Contoh HTTP Request (cURL):

curl -X POST https://api.firecrawl.dev/v2/scrape \
  -H "Authorization: Bearer fc-YOUR-API-KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://firecrawl.dev",
    "formats": ["markdown", "html"]
  }'

B. API Self-Hosted

Installation dengan Docker:

# Clone repository
git clone https://github.com/firecrawl/firecrawl.git
cd firecrawl

# Copy env file
cp apps/api/.env.example .env

# Edit .env - basic setup:
# USE_DB_AUTHENTICATION=false
# TEST_API_KEY=fc-my-firecrawl-key

# Build dan run
docker compose build
docker compose up -d

Akses API:

  • Base URL: http://localhost:3002
  • Gunakan TEST_API_KEY dari file .env
  • Semua endpoint v2 tersedia

Contoh ke API self-hosted:

curl -X POST http://localhost:3002/v2/scrape \
  -H "Authorization: Bearer fc-my-firecrawl-key" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "formats": ["markdown"]
  }'

3️⃣ Cara Akses MCP (Model Context Protocol)

Apa itu Firecrawl MCP?

Firecrawl MCP adalah server MCP yang memungkinkan AI model (seperti Claude, Cursor, ChatGPT) mengakses kemampuan web scraping Firecrawl melalui protokol JSON-RPC 2.0 yang distandarisasi oleh Anthropic.

Installation & Setup

Prerequisites:

  • Node.js dan npm
  • Firecrawl API key
  • MCP-compatible environment (Claude Desktop, Cursor IDE, dsb.)

Install MCP Server:

# Option 1: Jalankan langsung
FIRECRAWL_API_KEY=fc-yourkeyhere npx -y firecrawl-mcp

# Option 2: Install globally
npm install -g firecrawl-mcp
firecrawl-mcp

Environment Variables:

Variable Required Description
FIRECRAWL_API_KEY Yes API key Firecrawl
FIRECRAWL_API_URL Optional Custom endpoint untuk self-hosted (misal: https://firecrawl.yourdomain.com)

Setup di Cursor IDE

  1. Buka Cursor Settings
  2. Navigasi ke Features → MCP Servers
  3. Klik "+ Add New MCP Server"
  4. Masukkan konfigurasi:
    • Name: firecrawl-mcp
    • Type: command
    • Command: env FIRECRAWL_API_KEY=fc-your-api-key npx -y firecrawl-mcp

Setup di Claude Desktop

Tambahkan ke file konfigurasi Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": ["-y", "firecrawl-mcp"],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-api-key"
      }
    }
  }
}

Penggunaan MCP

Setelah terhubung, Anda dapat:

  • Meminta AI untuk mengambil konten dari website tertentu
  • Extract structured data dari web pages
  • Cari web content yang dirender dengan JavaScript

Contoh prompt di Claude/Cursor:

"Ambil judul dan konten utama dari halaman https://firecrawl.dev"

MCP server akan menangani rendering JavaScript, ekstraksi, dan mengembalikan data dalam format JSON yang siap untuk LLM.

📚 Ringkasan Singkat

Topik v1 v2
API URL /v1/ /v2/
SDK Class FirecrawlApp Firecrawl
Methods scrapeUrl(), mapUrl() scrape(), map()
Caching Manual Default 2 hari
Formats markdown, html + summary, json, screenshot objek
Crawl Options maxDepth maxDiscoveryDepth, prompt

Resources Tambahan:

Read next

Comments ()