Hosted scraping platforms compared: criteria, not a leaderboard

Hosted scraping platforms provide managed browsers, proxies, unblocking, or ready-made scrapers as a service. They can save operational time, but they also add a vendor dependency, a terms-of-service layer, and a cost model you need to understand before committing.

This is not a ranking. We have not run a controlled performance test against these vendors, and a leaderboard without a shared workload would be marketing. The table is a framework for your own evaluation, with links to the vendors’ documentation.

The categories are different

A “scraper marketplace” is not the same product as a “proxy network” or a “browser endpoint”. Compare vendors within a category first, then decide whether you need the category at all.

Criteria that matter

1. Access model

• Do you send a URL and get HTML/JSON back?
• Do you run code inside the platform?
• Do you get a browser session you control?
• Do you buy proxies separately?

A platform that returns parsed JSON is not interchangeable with one that gives you a raw browser. The integration work is different.

2. Rendering and JavaScript

• Does it render JavaScript by default, or only on request?
• Which browser engine and version?
• Can you wait for a selector, click, or scroll?
• Can you intercept network requests?

If your targets are server-rendered, you may not need a browser at all. Paying for rendering you do not use is a common waste.

3. Proxies and geo-targeting

• Which proxy types are included, and at what price?
• Can you choose a country or city?
• Is the proxy pool shared or dedicated?
• What is the acceptable-use policy?

Residential and mobile proxies raise consent and privacy questions. Read how the vendor sources them. “We have millions of IPs” is not a sourcing policy.

4. Pricing transparency

• Is there a free tier or trial?
• Is billing per request, per GB, per browser-minute, or per result?
• Are failures billed?
• Is there a minimum commit?
• Do prices change with volume, and is the current price easy to find?

Prices change. We deliberately do not copy a price table here because a stale number is worse than a link. Check the vendor’s pricing page on the date you evaluate it and record that date in your own notes.

5. Terms and compliance

• Does the vendor’s acceptable-use policy match your use case?
• Does it prohibit the sites or data categories you need?
• Who is responsible for the target site’s terms: you or the vendor?
• Does it offer a data-processing agreement for personal data?
• Where is data processed and stored?

A platform cannot transfer permission you do not have. If your crawl would breach a site’s terms, buying a proxy or a scraping API does not fix that.

6. Reliability and support

• What is the documented uptime or SLA?
• What happens when a request fails: retry, refund, or charge?
• Is there a status page and an incident history?
• Can you export your data and leave?

A vendor with no export path is a lock-in decision, not a tool choice.

A scorecard you can fill in

Use your own weights. A team that needs a managed browser will weight differently from a team that needs a JSON API for a public dataset.

When a hosted platform is the right choice

• You need access infrastructure you do not want to operate.
• Your volume is bursty and a managed pool is cheaper than idle capacity.
• You need geo-distributed requests and a vendor already has the coverage.
• You have read the terms and the use is permitted.

When it is not

• The data is available through an official API.
• The volume is small enough for a polite local crawler.
• The vendor’s terms prohibit your use case.
• You cannot explain the cost model or the data path.
• You need to own the data pipeline and the vendor has no export.

Sources

• Apify documentation
• Zyte documentation
• Bright Data documentation
• ScraperAPI documentation
• ScrapingBee documentation
• Browserless documentation