chore(infra): add separate browser service

Dockerfile.backend

@@ -18,9 +18,6 @@ COPY server/tsconfig.json ./server/
 # Install dependencies
 RUN npm install --legacy-peer-deps
 
-# Install Playwright browsers and dependencies
-RUN npx playwright install --with-deps chromium
-
 # Create the Chromium data directory with necessary permissions
 RUN mkdir -p /tmp/chromium-data-dir && \
     chmod -R 777 /tmp/chromium-data-dir

ENVEXAMPLE

@@ -38,3 +38,8 @@ AIRTABLE_REDIRECT_URI=http://localhost:8080/auth/airtable/callback
 
 # Telemetry Settings - Please keep it enabled. Keeping it enabled helps us understand how the product is used and assess the impact of any new changes. 
 MAXUN_TELEMETRY=true
+
+# WebSocket port for browser CDP connections
+BROWSER_WS_PORT=3001
+BROWSER_HEALTH_PORT=3002
+BROWSER_WS_HOST=browser

browser/.dockerignore

@@ -0,0 +1,9 @@
+node_modules
+npm-debug.log
+.env
+.git
+.gitignore
+dist
+*.ts
+!*.d.ts
+tsconfig.json

browser/Dockerfile

@@ -0,0 +1,30 @@
+FROM mcr.microsoft.com/playwright:v1.57.0-jammy
+
+WORKDIR /app
+
+# Copy package files
+COPY browser/package*.json ./
+
+# Install dependencies
+RUN npm ci
+
+# Copy TypeScript source and config
+COPY browser/server.ts ./
+COPY browser/tsconfig.json ./
+
+# Build TypeScript
+RUN npm run build
+
+# Accept build arguments for ports (with defaults)
+ARG BROWSER_WS_PORT=3001
+ARG BROWSER_HEALTH_PORT=3002
+
+# Set as environment variables
+ENV BROWSER_WS_PORT=${BROWSER_WS_PORT}
+ENV BROWSER_HEALTH_PORT=${BROWSER_HEALTH_PORT}
+
+# Expose ports dynamically based on build args
+EXPOSE ${BROWSER_WS_PORT} ${BROWSER_HEALTH_PORT}
+
+# Start the browser service (run compiled JS)
+CMD ["node", "dist/server.js"]

browser/package.json

@@ -0,0 +1,21 @@
+{
+    "name": "maxun-browser-service",
+    "version": "1.0.0",
+    "description": "Browser service that exposes Playwright browsers via WebSocket with stealth plugins",
+    "main": "dist/server.js",
+    "scripts": {
+        "build": "tsc",
+        "start": "node dist/server.js",
+        "dev": "ts-node server.ts"
+    },
+    "dependencies": {
+        "playwright": "1.57.0",
+        "playwright-extra": "^4.3.6",
+        "puppeteer-extra-plugin-stealth": "^2.11.2"
+    },
+    "devDependencies": {
+        "@types/node": "^22.7.9",
+        "typescript": "^5.0.0",
+        "ts-node": "^10.9.2"
+    }
+}

browser/server.ts

@@ -0,0 +1,92 @@
+import { chromium } from 'playwright-extra';
+import stealthPlugin from 'puppeteer-extra-plugin-stealth';
+import http from 'http';
+import type { BrowserServer } from 'playwright';
+
+// Apply stealth plugin to chromium
+chromium.use(stealthPlugin());
+
+let browserServer: BrowserServer | null = null;
+
+// Configurable ports with defaults
+const BROWSER_WS_PORT = parseInt(process.env.BROWSER_WS_PORT || '3001', 10);
+const BROWSER_HEALTH_PORT = parseInt(process.env.BROWSER_HEALTH_PORT || '3002', 10);
+
+async function start(): Promise<void> {
+    console.log('Starting Maxun Browser Service...');
+    console.log(`WebSocket port: ${BROWSER_WS_PORT}`);
+    console.log(`Health check port: ${BROWSER_HEALTH_PORT}`);
+
+    try {
+        // Launch browser server that exposes WebSocket endpoint
+        browserServer = await chromium.launchServer({
+            headless: true,
+            args: [
+                '--disable-blink-features=AutomationControlled',
+                '--disable-web-security',
+                '--disable-features=IsolateOrigins,site-per-process',
+                '--disable-site-isolation-trials',
+                '--disable-extensions',
+                '--no-sandbox',
+                '--disable-dev-shm-usage',
+                '--disable-gpu',
+                '--force-color-profile=srgb',
+                '--force-device-scale-factor=2',
+                '--ignore-certificate-errors',
+                '--mute-audio'
+            ],
+            port: BROWSER_WS_PORT,
+        });
+
+        console.log(`✅ Browser WebSocket endpoint ready: ${browserServer.wsEndpoint()}`);
+        console.log(`✅ Stealth plugin enabled`);
+
+        // Health check HTTP server
+        const healthServer = http.createServer((req, res) => {
+            if (req.url === '/health') {
+                res.writeHead(200, { 'Content-Type': 'application/json' });
+                res.end(JSON.stringify({
+                    status: 'healthy',
+                    wsEndpoint: browserServer?.wsEndpoint(),
+                    wsPort: BROWSER_WS_PORT,
+                    healthPort: BROWSER_HEALTH_PORT,
+                    timestamp: new Date().toISOString()
+                }));
+            } else if (req.url === '/') {
+                res.writeHead(200, { 'Content-Type': 'text/plain' });
+                res.end(`Maxun Browser Service\nWebSocket: ${browserServer?.wsEndpoint()}\nHealth: http://localhost:${BROWSER_HEALTH_PORT}/health`);
+            } else {
+                res.writeHead(404);
+                res.end('Not Found');
+            }
+        });
+
+        healthServer.listen(BROWSER_HEALTH_PORT, () => {
+            console.log(`✅ Health check server running on port ${BROWSER_HEALTH_PORT}`);
+            console.log('Browser service is ready to accept connections!');
+        });
+    } catch (error) {
+        console.error('❌ Failed to start browser service:', error);
+        process.exit(1);
+    }
+}
+
+// Graceful shutdown
+async function shutdown(): Promise<void> {
+    console.log('Shutting down browser service...');
+    if (browserServer) {
+        try {
+            await browserServer.close();
+            console.log('Browser server closed');
+        } catch (error) {
+            console.error('Error closing browser server:', error);
+        }
+    }
+    process.exit(0);
+}
+
+process.on('SIGTERM', shutdown);
+process.on('SIGINT', shutdown);
+
+// Start the service
+start().catch(console.error);

browser/tsconfig.json

@@ -0,0 +1,24 @@
+{
+    "compilerOptions": {
+        "target": "ES2020",
+        "module": "commonjs",
+        "lib": [
+            "ES2020"
+        ],
+        "outDir": "./dist",
+        "rootDir": "./",
+        "strict": true,
+        "esModuleInterop": true,
+        "skipLibCheck": true,
+        "forceConsistentCasingInFileNames": true,
+        "resolveJsonModule": true,
+        "moduleResolution": "node"
+    },
+    "include": [
+        "server.ts"
+    ],
+    "exclude": [
+        "node_modules",
+        "dist"
+    ]
+}

docker-compose.yml

@@ -74,6 +74,42 @@ services:
     depends_on:
       - backend
 
+  browser:
+    build:
+      context: .
+      dockerfile: browser/Dockerfile
+      args:
+        BROWSER_WS_PORT: ${BROWSER_WS_PORT:-3001}
+        BROWSER_HEALTH_PORT: ${BROWSER_HEALTH_PORT:-3002}
+    ports:
+      - "${BROWSER_WS_PORT:-3001}:${BROWSER_WS_PORT:-3001}"
+      - "${BROWSER_HEALTH_PORT:-3002}:${BROWSER_HEALTH_PORT:-3002}"
+    environment:
+      - NODE_ENV=production
+      - DEBUG=pw:browser*
+      - BROWSER_WS_PORT=${BROWSER_WS_PORT:-3001}
+      - BROWSER_HEALTH_PORT=${BROWSER_HEALTH_PORT:-3002}
+    restart: unless-stopped
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:${BROWSER_HEALTH_PORT:-3002}/health"]
+      interval: 10s
+      timeout: 5s
+      retries: 3
+      start_period: 10s
+    deploy:
+      resources:
+        limits:
+          memory: 2G
+          cpus: '1.5'
+        reservations:
+          memory: 1G
+          cpus: '1.0'
+    security_opt:
+      - seccomp:unconfined
+    shm_size: 2gb
+    cap_add:
+      - SYS_ADMIN
+
 volumes:
   postgres_data:
   minio_data:

maxun-core/package.json

@@ -31,10 +31,10 @@
   "license": "AGPL-3.0-or-later",
   "dependencies": {
     "@cliqz/adblocker-playwright": "^1.31.3",
+    "@types/node": "22.7.9",
     "cross-fetch": "^4.0.0",
     "joi": "^17.6.0",
-    "playwright": "^1.20.1",
-    "playwright-extra": "^4.3.6",
-    "puppeteer-extra-plugin-stealth": "^2.11.2"
+    "playwright-core": "1.57.0",
+    "turndown": "^7.2.2"
   }
-}
+}

maxun-core/src/interpret.ts

@@ -1,5 +1,5 @@
 /* eslint-disable no-await-in-loop, no-restricted-syntax */
-import { ElementHandle, Page, PageScreenshotOptions } from 'playwright';
+import { ElementHandle, Page, PageScreenshotOptions } from 'playwright-core';
 import { PlaywrightBlocker } from '@cliqz/adblocker-playwright';
 import fetch from 'cross-fetch';
 import path from 'path';
@@ -144,7 +144,7 @@ export default class Interpreter extends EventEmitter {
   private async applyAdBlocker(page: Page): Promise<void> {
     if (this.blocker) {
       try {
-        await this.blocker.enableBlockingInPage(page);
+        await this.blocker.enableBlockingInPage(page as any);
       } catch (err) {
         this.log(`Ad-blocker operation failed:`, Level.ERROR);
       }
@@ -154,7 +154,7 @@ export default class Interpreter extends EventEmitter {
   private async disableAdBlocker(page: Page): Promise<void> {
     if (this.blocker) {
       try {
-        await this.blocker.disableBlockingInPage(page);
+        await this.blocker.disableBlockingInPage(page as any);
       } catch (err) {
         this.log(`Ad-blocker operation failed:`, Level.ERROR);
       }

maxun-core/src/types/workflow.ts

@@ -1,4 +1,4 @@
-import { Page } from 'playwright';
+import { Page } from 'playwright-core';
 import {
   naryOperators, unaryOperators, operators, meta,
 } from './logic';

package.json

@@ -55,10 +55,9 @@
     "node-cron": "^3.0.3",
     "pg": "^8.13.0",
     "pg-boss": "^10.1.6",
-    "playwright": "^1.48.2",
-    "playwright-extra": "^4.3.6",
+    "pkce-challenge": "^4.1.0",
+    "playwright-core": "1.57.0",
     "posthog-node": "^4.2.1",
-    "puppeteer-extra-plugin-stealth": "^2.11.2",
     "react": "^18.0.0",
     "react-dom": "^18.0.0",
     "react-i18next": "^15.1.3",
@@ -129,4 +128,4 @@
     "vite": "^5.4.10",
     "zod": "^3.25.62"
   }
-}
+}

server/src/api/record.ts

@@ -1,6 +1,4 @@
 import { Router, Request, Response } from 'express';
-import { chromium } from "playwright-extra";
-import stealthPlugin from 'puppeteer-extra-plugin-stealth';
 import { requireAPIKey } from "../middlewares/api";
 import Robot from "../models/Robot";
 import Run from "../models/Run";
@@ -20,8 +18,6 @@ import { airtableUpdateTasks, processAirtableUpdates } from "../workflow-managem
 import { sendWebhook } from "../routes/webhook";
 import { convertPageToHTML, convertPageToMarkdown } from '../markdownify/scrape';
 
-chromium.use(stealthPlugin());
-
 const router = Router();
 
 const formatRecording = (recordingData: any) => {