Ideas Weave Every Narrative with AI.
AI companies are scraping websites for free, forcing small site owners to pay for their bandwidth and compute. This hidden infrastructure tax is reshaping the web, locking out independent voices. The real battle isn’t over copyright—it’s about who pays for the servers.
Fitter is a declarative web-data extraction tool that runs entirely in the browser—no servers, no API keys, no proxy rotation. It exposes an uncomfortable truth: most scraping infrastructure exists because developers never questioned the default. The browser already does everything a scraping backend does, but better and for free.
Anubis uses proof-of-work to block AI scrapers, but it doesn’t stop well-funded AI companies — it only blocks independent developers, open-source tools, and hobbyists. The tool accelerates the death of the open web by handing a monopoly on data to the very entities it claims to fight.
AI agents are wasting massive compute resources by relying on full-fledged browsers like Chrome. Ember, a 17MB headless browser, strips away the unnecessary GUI overhead to let agents parse web content directly. This isn’t just a tweak — it’s a fundamental shift in how we should design agent infrastructure, cutting memory usage by 90% and halving cloud costs.