Imported from jeremylongshore/tons-of-skills-marketplace (plugins/saas-packs/brightdata-pack/skills/brightdata-reference-architecture/SKILL.md). Install upstream with npx skills add jeremylongshore/tons-of-skills-marketplace --skill brightdata-reference-architecture. Copyright stays with the author (MIT).

Bright Data Reference Architecture

Overview

Production-ready architecture for Bright Data scraping systems. Covers project layout, data pipeline design, and integration patterns for Web Unlocker, Scraping Browser, SERP API, and Datasets API.

Prerequisites

• Understanding of layered architecture
• Node.js/TypeScript project setup
• Database for storing scraped data

Authentication

Resolve the zone credential only from the approved server-side secret manager at runtime. Do not expose it to browsers, logs, source control, job payloads, or support bundles. Associate every credential with an owner, approved target/data policy, rotation schedule, and revocation path.

Instructions

1. Accept a collection request only after validating caller authorization, target allowlist, legal/policy purpose, data class, and retention requirement.
2. Create a durable idempotency record and apply rate/cost limits before dispatching the validated request to the proxy adapter.
3. Store minimum necessary results in protected storage and run schema/data-class checks before downstream processing.
4. Emit redacted telemetry, reconcile completion, and pause the queue on target-policy, authorization, or unexpected-data failures.

Project Structure

my-scraper/
├── src/
│   ├── brightdata/
│   │   ├── proxy.ts            # Proxy config helper (zone, country, session)
│   │   ├── client.ts           # Axios client with proxy + retry
│   │   ├── browser.ts          # Scraping Browser connection manager
│   │   ├── api.ts              # REST API client (trigger, snapshot)
│   │   ├── cache.ts            # Response cache (LRU + optional Redis)
│   │   └── types.ts            # Shared TypeScript interfaces
│   ├── scrapers/
│   │   ├── product-scraper.ts  # Domain-specific scraper
│   │   ├── serp-scraper.ts     # Search result collector
│   │   └── parser.ts           # HTML → structured data (cheerio)
│   ├── pipeline/
│   │   ├── scheduler.ts        # Cron-based scraping scheduler
│   │   ├── processor.ts        # Raw HTML → clean data
│   │   └── storage.ts          # Database/file output
│   ├── webhooks/
│   │   └── brightdata.ts       # Webhook delivery handler
│   └── api/
│       ├── health.ts           # Health check endpoint
│       └── scrape.ts           # On-demand scrape endpoint
├── tests/
│   ├── unit/                   # Mocked tests (no proxy needed)
│   ├── integration/            # Live proxy tests
│   └── fixtures/               # Cached HTML for testing
├── config/
│   ├── zones.json              # Zone configuration per environment
│   └── targets.json            # Target URLs and scraping schedules
└── .env.example

Architecture Diagram

┌──────────────────────────────────────────────────────┐
│                    API / Scheduler                     │
│         (On-demand scrape, cron jobs, webhooks)        │
├──────────────────────────────────────────────────────┤
│                   Scraper Layer                        │
│    (Product scraper, SERP scraper, custom parsers)     │
├────────────┬─────────────────┬───────────────────────┤
│ Web        │  Scraping       │  SERP / Datasets      │
│ Unlocker   │  Browser        │  API                  │
│ (Proxy)    │  (WebSocket)    │  (REST)               │
├────────────┴─────────────────┴───────────────────────┤
│          Bright Data Infrastructure Layer              │
│  (Proxy config, retry, cache, session management)      │
├──────────────────────────────────────────────────────┤
│              Storage / Pipeline                        │
│    (Database, file output, webhook delivery)           │
└──────────────────────────────────────────────────────┘

Key Components

Step 1: Multi-Product Client

// src/brightdata/client.ts
import axios, { AxiosInstance } from 'axios';
import https from 'https';
import { chromium } from 'playwright';

export class BrightDataClient {
  private proxyClient: AxiosInstance;
  private apiToken: string;

  constructor(private config: {
    customerId: string;
    zone: string;
    zonePassword: string;
    apiToken: string;
  }) {
    this.apiToken = config.apiToken;
    this.proxyClient = axios.create({
      proxy: {
        host: 'brd.superproxy.io',
        port: 33335,
        auth: {
          username: `brd-customer-${config.customerId}-zone-${config.zone}`,
          password: config.zonePassword,
        },
      },
      httpsAgent: new https.Agent({ keepAlive: true, rejectUnauthorized: false }),
      timeout: 60000,
    });
  }

  // Web Unlocker — simple HTTP through proxy
  async scrape(url: string, country?: string): Promise<string> {
    const response = await this.proxyClient.get(url);
    return response.data;
  }

  // Scraping Browser — Playwright over CDP
  async scrapeWithBrowser(url: string, extract: (page: any) => Promise<any>) {
    const auth = `brd-customer-${this.config.customerId}-zone-scraping_browser1:${this.config.zonePassword}`;
    const browser = await chromium.connectOverCDP(`wss://${auth}@brd.superproxy.io:9222`);
    try {
      const page = await browser.newPage();
      await page.goto(url, { waitUntil: 'domcontentloaded', timeout: 60000 });
      return await extract(page);
    } finally {
      await browser.close();
    }
  }

  // Web Scraper API — async bulk collection
  async triggerCollection(datasetId: string, inputs: any[]) {
    const response = await fetch(
      `https://api.brightdata.com/datasets/v3/trigger?dataset_id=${datasetId}&format=json`,
      {
        method: 'POST',
        headers: { 'Authorization': `Bearer ${this.apiToken}`, 'Content-Type': 'application/json' },
        body: JSON.stringify(inputs),
      }
    );
    return response.json();
  }
}

Step 2: Scraping Pipeline

// src/pipeline/scheduler.ts
import cron from 'node-cron';

interface ScrapeJob {
  name: string;
  urls: string[];
  product: 'web_unlocker' | 'scraping_browser' | 'datasets_api';
  schedule: string; // cron expression
  parser: (html: string) => any;
}

export function startScheduler(jobs: ScrapeJob[], client: BrightDataClient) {
  for (const job of jobs) {
    cron.schedule(job.schedule, async () => {
      console.log(`Running job: ${job.name}`);
      if (job.product === 'datasets_api') {
        await client.triggerCollection('dataset_id', job.urls.map(url => ({ url })));
      } else {
        for (const url of job.urls) {
          const html = await client.scrape(url);
          const data = job.parser(html);
          await saveToDatabase(job.name, data);
        }
      }
    });
  }
}

Step 3: Environment Configuration

// config/zones.json
{
  "development": {
    "web_unlocker": "web_unlocker_dev",
    "scraping_browser": "scraping_browser_dev",
    "api_datasets": true
  },
  "production": {
    "web_unlocker": "web_unlocker_prod",
    "scraping_browser": "scraping_browser_prod",
    "api_datasets": true
  }
}

Decision Matrix

Output

• Multi-product Bright Data client
• Domain-specific scrapers with parsers
• Cron-based scraping pipeline
• Environment-isolated zone configuration

Examples

Place target authorization, request validation, rate/budget controls, and audit logging ahead of the proxy adapter. The queue persists an idempotency key, workers validate the target policy before every call, and protected storage retains only minimum necessary data. An unexpected content class pauses the job instead of flowing to downstream systems.

Error Handling

Resources

• Bright Data Products Overview
• Scraping Browser
• Web Scraper API
• SERP API

Next Steps

For multi-environment setup, see brightdata-deploy-integration.