How Nous Research used BotID to block automated abuse at scale

Nous Research, an AI lab, temporarily made its open-source language model, Hermes, free to enhance accessibility, but the initiative was exploited by automated scripts that overwhelmed the service with fake accounts to bypass rate limits, leading to wasted resources and inflated costs. Despite having initial bot protection measures like Cloudflare Turnstile, these were insufficient to prevent abuse during the promotional period. In response, Nous implemented Vercel BotID Deep Analysis for enhanced bot detection and prevention, which operates invisibly and verifies users without disrupting their experience. This new system enabled Nous to effectively block a coordinated attack that increased traffic by 3,000% without affecting legitimate users or service availability, allowing them to continue offering the free tier safely. The improved bot protection not only safeguarded their infrastructure but also maintained cost efficiency and app performance, highlighting the importance of robust security measures for open-access platforms.