DEV Community

Session zero
Session zero

Posted on

From Apify to Everywhere: Building Multi-Channel Distribution for Korean Data Scrapers

After hitting 5,500+ runs and 71 users on Apify, I realized something: Apify is great, but it's only one channel. Different developers live in different ecosystems. n8n users want drag-and-drop nodes. API consumers want REST endpoints on RapidAPI. Neither group is browsing the Apify Store.

So in 48 hours, I built distribution for both — 3 n8n community nodes and 3 Cloudflare Worker proxies for RapidAPI. Here's how.

Why Multi-Channel?

My Korean data scrapers (Naver News, Naver Place, Melon Chart) were only accessible through Apify's platform. That's a problem:

  • n8n/Zapier users want native integrations, not custom HTTP nodes
  • Mobile/SaaS developers want simple REST APIs, not actor run management
  • Enterprise teams may already have RapidAPI subscriptions

Same data, different packaging. Each channel reaches users who'd never find you on Apify.

Channel 1: n8n Community Nodes

n8n is the open-source workflow automation platform with 400K+ users. Their community node system lets you publish custom integrations as npm packages.

What I Built

Three npm packages, each wrapping one Apify actor:

Package Actor Runs
n8n-nodes-naver-news-scraper Naver News Scraper 3,283
n8n-nodes-naver-place-search Naver Place Search 695
n8n-nodes-melon-chart-scraper Melon Chart Scraper 26

How It Works

Apify provides an official n8n template that reads your actor's input schema and auto-generates n8n node parameters. The process:

  1. Fork the template
  2. Point it at your actor ID
  3. Fix TypeScript compilation errors (there were a few — noUnusedLocals and apostrophe escaping in descriptions)
  4. Build and publish to npm

Each node supports both API key and OAuth2 authentication, and includes AI tool integration for n8n's AI agent workflows.

Bonus: Workflow Templates

I also created ready-to-import workflow templates:

  • Naver News → Google Sheets (keyword monitoring)
  • Melon Chart → Discord (K-pop chart alerts)
  • Naver Place Reviews → Email Digest

These lower the barrier from "install the node" to "import and run."

GitHub repos:

Channel 2: RapidAPI via Cloudflare Workers

RapidAPI is the world's largest API marketplace. The catch? You need actual REST endpoints — Apify's actor-based model doesn't map directly.

The Architecture

RapidAPI → Cloudflare Worker → Apify Actor (sync run) → JSON response
Enter fullscreen mode Exit fullscreen mode

Each Cloudflare Worker:

  • Validates X-RapidAPI-Proxy-Secret for security
  • Maps clean REST endpoints to Apify actor inputs
  • Calls run-sync-get-dataset-items for synchronous execution
  • Returns structured JSON with CORS support

Endpoints

Naver News API:

  • GET /search?query=삼성&maxItems=10 — keyword search
  • GET /section?section=politics — section browsing
  • GET /ranking?type=mostViewed — trending articles

Naver Place API:

  • GET /search?query=강남+맛집&maxItems=20 — place search with reviews

Melon Chart API:

  • GET /chart?chartType=TOP100 — real-time music charts
  • GET /search?query=IVE — artist/song search

Cost Structure

Layer Cost
Cloudflare Workers Free (100K req/day)
Apify Pay-per-event (actor-specific)
RapidAPI 25% commission
Net margin ~50-60%

The zero-cost infrastructure layer is key — Cloudflare's free tier handles the proxy, so you only pay for actual Apify compute.

The Blue Ocean

I searched RapidAPI for Korean data APIs. The results? Almost nothing. One Naver Shop scraper. Zero for Naver News, Naver Place, or Melon Chart. This is a completely uncontested market.

Live Worker URLs:

GitHub repos:

The Distribution Matrix

After one week, here's the full picture:

Channel Type Status Reach
Apify Store Actor marketplace ✅ Live (5,500+ runs) Apify users
npm / n8n Community nodes ✅ Built (npm pending) 400K+ n8n users
RapidAPI REST API marketplace ✅ Built (listing pending) 4M+ developers
Dev.to Content marketing ✅ 16 articles SEO + dev community
MCP Server AI tool protocol ✅ Published AI agent developers
PyPI Python package ✅ Published Python developers

Same 13 actors. Six distribution channels. Each one compounds — an n8n user might discover the Apify actor, a Dev.to reader might try the RapidAPI endpoint.

What I Learned

  1. Build once, distribute everywhere. The hardest part was building the scrapers. Wrapping them for different platforms took hours, not weeks.
  2. Each platform has its own discovery. n8n has a community forum, RapidAPI has search, Dev.to has SEO. You're invisible on platforms you're not on.
  3. Cloudflare Workers are perfect API proxies. Free tier, TypeScript, global edge deployment — ideal for wrapping any backend.
  4. Uncontested markets exist. Korean data on RapidAPI is virtually empty. Sometimes the opportunity is just showing up.

Building Korean data tools and sharing the journey. Previous posts: 5,500 runs milestone | First revenue | All Korean scraper guides

Top comments (0)