Zyte (formerly Scrapinghub) is an enterprise scraping platform. WebPerception is a single API that gives your code eyes. Here's how they differ.
| Feature | WebPerception API | Zyte |
|---|---|---|
| AI Data Extraction | ✅ Schema-based, any page | ✅ Zyte API (automatic extraction) |
| Setup Complexity | ✅ 1 API call, 2 minutes | ❌ Multiple products, steep learning curve |
| Self-Serve Pricing | ✅ Public pricing, instant signup | ⚠️ "Contact sales" for most plans |
| JavaScript Rendering | ✅ Full browser | ✅ Full browser |
| Screenshots | ✅ Built-in | ⚠️ Via Scrapy spiders |
| Proxy Infrastructure | ✅ Built-in (transparent) | ✅ Smart Proxy Manager (65M+ IPs) |
| Agent Integrations | ✅ LangChain, CrewAI, AutoGen | ❌ Not agent-focused |
| Free Tier | 100 calls/month | 14-day trial |
| Starting Price | $29/mo (5,000 calls) | ~$300+/mo (enterprise) |
| Target User | Developers, agent builders | Enterprise data teams |
With Zyte (Scrapy Cloud):
# 1. Write a Scrapy spider
import scrapy
class ProductSpider(scrapy.Spider):
name = 'products'
start_urls = ['https://example.com/product']
def parse(self, response):
yield {
'title': response.css('h1.product-title::text').get(),
'price': response.css('.price::text').get(),
'stock': response.css('.availability::text').get(),
}
# 2. Deploy to Scrapy Cloud
# 3. Configure proxy settings
# 4. Schedule crawl job
# 5. Download results from storageWith WebPerception:
import requests
response = requests.post("https://api.mantisapi.com/extract", json={
"url": "https://example.com/product",
"schema": {
"title": "string",
"price": "number",
"currency": "string",
"in_stock": "boolean"
}
})
product = response.json()["data"]
# Done. No spiders. No deployment. No proxy config.Choose Zyte if: You're an enterprise with dedicated data engineering teams, need massive-scale recurring crawls, and want a full scraping infrastructure (proxies, Scrapy hosting, storage). Zyte is the established enterprise player.
Choose WebPerception if: You want to extract data from web pages without building crawl infrastructure. Especially suited for AI agent developers who need real-time web perception — not batch crawling jobs. Our API is simpler, cheaper, and purpose-built for the agent era.
Bottom line: Zyte is built for data engineering teams running large crawl operations. WebPerception is built for developers who just need their code to understand web pages. Different tools for different jobs — but if simplicity and AI agents are your priority, we're the clear choice.