TRACE: a daily, open-source benchmark of web scraping APIs
Blog post from Context.dev
TRACE is a daily, open-source web-scraping benchmark created by Context.dev, which also competes in it, designed to measure providers on newly discovered pages rather than static test sets. It samples up to 300 previously untested pages per category from 197 feeds, registries, sitemaps, and Wikipedia links, then sends uncached requests to Context.dev, Keenable, Firecrawl, and Exa, judging blinded responses for substantive content, speed, and estimated cost. In its November 6, 2026 sample of 2,527 pages, Context.dev led with an 87.0% useful-content rate, ahead of Keenable at 84.2%, Firecrawl at 82.3%, and Exa at 79.1%, while also reporting the shortest p95 latency at 8.91 seconds and an estimated cost of $0.79 per 1,000 useful pages. TRACE publishes its code, prompts, settings, and page-level records to support reproduction and scrutiny, but acknowledges limitations including potential judge errors, source-selection bias, unresolved results counted as failures, measurements from one region, and the inherent conflict of a benchmark operator ranking its own product.
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.