Technical notes on web development, DevOps, and AI integration.
1 article
Finding a site's own GraphQL endpoint in its public JS bundle: 5330 rows in 7 requests, not 533 HTML pages.
TL;DR: Expecting a tedious HTML scraping job for Indonesia's procurement blacklist, the author found an unauthenticated GraphQL endpoint exposed by the site's XLSX export. A perPage of 800 meant the full 5,330-row dataset downloaded in just seven requests and about ten seconds. The daily snapshot plan was retired in favor of a simple full-dataset sync.