Automating Kryfto with n8n

February 23, 2026 ยท View on GitHub

Kryfto is designed from the ground up to operate as an asynchronous, webhook-compatible headless browser engine. This makes it the perfect companion for visual workflow automation tools like n8n.

Unlike basic scraping widgets, Kryfto runs full Chromium browsers capable of executing JavaScript, bypassing Cloudflare/Datadome (via stealth mode), and returning structured JSON.

๐Ÿ—๏ธ Core n8n Architecture

The most reliable way to integrate Kryfto into n8n is using the HTTP Request Node.

Since advanced website scraping (waiting for selectors, clicking pagination) takes time, Kryfto uses a Job-Queue architecture. Rather than holding the HTTP request open for 60 seconds and risking a timeout, you can instruct Kryfto to execute the job and instantly POST the results back to an n8n Webhook Node.

The Webhook Polling Payload

When creating a job, you can use the webhook parameter. Kryfto will ping this URL when the Playwright browser session finishes.

{
  "url": "https://news.ycombinator.com",
  "webhook": "https://your-n8n-instance.com/webhook/my-kryfto-listener",
  "options": {
    "browserEngine": "chromium"
  },
  "extract": {
    "mode": "selectors",
    "selectors": {
      "topStories": ".titleline > a"
    }
  }
}

๐Ÿš€ Scenario: The Daily Price Monitor Dashboard

Goal: Every morning at 8:00 AM, scrape a list of competitor eCommerce products and insert the prices into a Google Sheet.

Step-by-Step n8n Pipeline:

1. Setup the Trigger

  • Add a Schedule Trigger node.
  • Configure it to run daily at 08:00.

2. Dispatch to Kryfto (HTTP Request Node)

  • Add an HTTP Request node.
  • Method: POST
  • URL: http://<your-kryfto-server-ip>:8080/v1/jobs
  • Authentication: Header Auth
    • Name: Authorization
    • Value: Bearer <KRYFTO_API_TOKEN>
  • Body Content Type: JSON
  • Parameters (Body):
{
  "url": "https://competitor.com/product/123",
  "wait": true,
  "options": {
    "timeoutMs": 30000
  },
  "extract": {
    "mode": "selectors",
    "selectors": {
      "productName": "h1.product-title",
      "price": "span.price-amount",
      "stockStatus": "div.inventory-level"
    }
  }
}

(Note: Setting "wait": true tells the Kryfto API to hold the connection open until the scrape finishes and return the final data immediately. This is simpler than webhooks for fast jobs under 30 seconds).

3. Parse and Insert (Google Sheets Node)

  • Add a Google Sheets node.
  • Operation: Append Row
  • Map the JSON response from Kryfto directly into your sheet columns:
    • Name Column: ={{ $json.data.extract.productName }}
    • Price Column: ={{ $json.data.extract.price }}

๐Ÿฅท Bypassing Bot Protection in n8n

If your n8n workflow hits a site that heavily protects itself (like Amazon or Cloudflare), you just need to ensure Kryfto's backend environment is configured properly.

You do not need to change your n8n nodes. Simply go to your Kryfto .env file and enable:

KRYFTO_STEALTH_MODE=true
KRYFTO_PROXY_URLS=socks5://proxy1:1080

This forces Kryfto to automatically inject anti-fingerprinting scripts and route the Playwright browsers through your residential proxy before loading the page. n8n will still receive the clean JSON!

๏ฟฝ Native Search Automation

You don't just have to scrape URLs you already know. Kryfto has a native /v1/search endpoint that queries search engines directly (without Google/Bing API limits).

You can use this inside an n8n HTTP Request node to find domains before you crawl them.

n8n HTTP Node Payload:

  • Method: POST
  • URL: http://<your-kryfto-server-ip>:8080/v1/search
  • Body:
{
  "query": "best enterprise headless CMS tools 2025",
  "limit": 5,
  "engine": "duckduckgo",
  "safeSearch": "moderate",
  "locale": "us-en"
}

n8n Chaining Pattern: You can route the JSON results array from this search node directly into an n8n Split In Batches Node, and then fire a separate Kryfto /v1/jobs crawling request for every single search result returned!

๏ฟฝ๐Ÿ”Ž Extracting Data via LLM (Schema Mode)

If the HTML on a page is too messy for CSS selectors, you can use Kryfto's LLM Schema Extraction. Kryfto will read the page and use an AI model to structure the data exactly how you want it.

n8n HTTP Node Payload:

{
  "url": "https://en.wikipedia.org/wiki/Apple_Inc.",
  "wait": true,
  "extract": {
    "mode": "schema",
    "schema": {
      "type": "object",
      "properties": {
        "companyName": { "type": "string" },
        "founders": {
          "type": "array",
          "items": { "type": "string" }
        },
        "revenue": { "type": "string", "description": "Latest annual revenue" }
      }
    }
  }
}

(Note: LLM extraction requires you to have OPENAI_API_KEY configured in Kryfto's .env file).