Bluefield
Choosing a web scraping API
The best web scraping API for your workload depends on coverage on your URL types, latency, credit cost, and whether you need provenance or watch mode. Benchmark figures are held until the latest competitive re-run is reviewed.
How to evaluate a web scraping API
Evaluate coverage on the page types you actually fetch, the credit cost per successful scrape, latency at p50 and p95, and whether the vendor publishes reproducible benchmarks.
- Coverage
- Share of URLs where the API returns usable extracted content.
- Stratified corpus
- A benchmark set grouped by page type such as docs, SPAs, and anti bot sites.
- Credit tariff
- Published credits charged per operation type.
Bluefield capabilities
| Capability | Detail |
|---|---|
| Measured benchmarks | Methodology and strata on /benchmarks |
| Credit based billing | Costs scale with fetch tier and add ons |
| Watch mode | Webhook on meaningful page changes |
| Provenance | Signed, verifiable captures on every live fetch |
| Self host | Docker deployment with the same API surface |
Frequently asked questions
- Does Bluefield publish benchmark numbers?
- The /benchmarks page describes our stratified URL corpus, method, and caveats. Headline figures are held until the latest re-run is reviewed.
- What page types are in the benchmark corpus?
- Twelve strata including clean articles, news, documentation, ecommerce, SPAs, anti bot hard sites, and multilingual pages.
- How much does a basic scrape cost?
- One credit for a static scrape. Browser and stealth tiers cost more. See /pricing for the full tariff.