TechPrufer
Get started
Vercel
Vercel logoVercel · Feature trail

Bot Detection and Mitigation

17 posts · May 2023 – May 2026

What this trail covers

Vercel's platform is evolving to address the growing challenge of AI bot traffic and Denial of Wallet (DoW) attacks. The platform offers tools like Vercel Firewall, Bot Protection, and BotID to differentiate between beneficial AI crawlers and malicious bots, and to protect against attacks that exploit cloud scalability to drive up operational costs. This includes granular API rate limiting, challenge-response protocols, behavioral analysis, and adaptive filtering to distinguish between human use.

All Vercel trails
  1. Protecting against token theftThis post introduces Vercel BotID with deep analysis as a per-request verification mechanism to prevent AI inference theft. It details the architecture of inference theft attacks, the limitations of traditional web defenses (IP rate limits, auth walls) against these attacks, and provides a concrete implementation example using `botid/server` and `botid/client` in a Next.js route handler. The core contribution is the per-request verification strategy, contrasting it with session-based checks, and highlighting the cost asymmetry between inference and verification.
  2. Security through design: Creating the improved Firewall experienceThis post details the redesign of the Vercel Firewall UI, focusing on improving user experience and visibility for developers, SREs, and security teams. Key technical contributions include: - **Redesigned Overview Page:** Consolidates key Firewall activity into four tables: Alerts (DDoS attacks), Rules (top rule activity), Events (mitigations), and Denied IPs. This provides a high-level control center for security posture. - **Comprehensive Traffic Intelligence:** The Traffic page now offers dedicated feeds for Top IPs, JA4 digests, AS names, User Agents, Request Paths, and active rules, enabling deeper analysis of traffic patterns and anomalies. - **Dedicated Rules and Activity Tabs:** Firewall Rules and the Audit Log are now accessible via dedicated sidebar navigation, simplifying management and visibility of WAF custom rules (Bot Protection, Managed Rulesets, IP Blocking) and changes. - **Faster Event Inspection:** Alerts and events now open detailed views directly within the page, allowing for immediate investigation of suspicious traffic or DDoS attacks without context switching.
  3. BotID Deep Analysis catches a sophisticated bot network in real-timeThis post details the real-time detection and adaptation of Vercel's BotID Deep Analysis system against a sophisticated new browser bot network. The system identified coordinated bot activity by correlating identical browser fingerprints appearing across multiple proxy nodes, a pattern that traditional security measures missed. The machine learning backend adapted within minutes, re-verifying sessions and classifying the traffic as malicious, leading to its blocking. This demonstrates the system's ability to learn and adapt to novel evasion techniques without manual intervention.
  4. BotID uncovers hidden SEO poisoningThis post details how BotID's bot identification capabilities helped uncover a legacy SEO poisoning attack by distinguishing legitimate Google crawlers from malicious traffic, enabling targeted mitigation strategies like updating robots.txt and submitting removal requests to Google Search Console.
  5. The three types of AI bot traffic and how to handle themThis post categorizes AI bot traffic into three types: AI training crawlers, AI engine grounding bots, and AI referrals. It explains the value of allowing AI crawlers for discoverability and potential referrals, likening it to early SEO. The post advises a selective blocking strategy, recommending blocking sensitive routes like `/login` and `/admin` while allowing access to discovery content such as documentation and product pages. It highlights Vercel Firewall, Bot Protection, and BotID as tools to manage this traffic.
  7. ​Introducing BotID, invisible bot filtering for critical routesIntroduces BotID, a new layer of protection on Vercel designed to stop sophisticated bots that mimic human behavior before they reach critical backend routes. BotID injects obfuscated, mutating JavaScript code to collect signals distinguishing humans from bots, feeding into a global ML network. It offers a simple server-side check (`checkBotId`) with no configuration required. Deep Analysis mode, powered by Kasada, provides enhanced protection for high-value applications.
  8. Vercel security roundup: improved bot defenses, DoS mitigations, and insightsThis post details the introduction of Bot Filter, a managed WAF ruleset for challenging non-browser bot traffic, and Protectd, an update to DoS mitigation infrastructure that significantly accelerates Layer 7 attack mitigation with a median mitigation time of 2.5 seconds. It also highlights proactive vulnerability management for open-source frameworks like React Router, Remix, and Next.js middleware, and introduces firewall usability improvements such as IP enrichment and the OR operator for rule conditions.
  9. Vercel's Bot Protection stops malicious bots in a single clickIntroduces the Bot Protection managed ruleset, a new security layer integrated with the Vercel WAF. This feature uses heuristics-based detection to distinguish between browser-based (human) and non-browser (bot) traffic with zero configuration. It allows requests from real browsers and verified bots (including project cron jobs and API calls from Vercel Functions) and can issue challenges to non-browser clients like `curl` or unwanted scrapers. The ruleset is maintained by Vercel, and a public beta feedback thread is available. The post also mentions future plans for deeper observability, more granular control over bots, an SDK, and runtime capabilities.
  10. Protectd: Evolving Vercel’s always-on denial-of-service mitigationsThis post introduces Protectd, Vercel's new real-time security engine for DoS mitigation. It details the evolution from a four-phase, polling-based system to an edge-based, stream-processing architecture using Golang. Key technical contributions include: moving detection to the edge for faster Time to Mitigation (TTM), replacing polling with direct event streams, using probabilistic data structures and JIT/SIMD for high-throughput event processing (~550K events/sec on 14 cores), implementing shadow mode and percentage-based rollouts for safe rule deployment, and enabling historical traffic replay for backtesting. The post quantifies the TTM improvement, achieving P99 of 3.5s and P50 of 2.5s, with response times as low as 0.5s.
  11. Vercel security roundup: Faster defenses and better visibility for your appsThis post details Vercel's security enhancements in the last quarter of 2024, focusing on WAF performance and visibility. Key contributions include: 1. Mitigation of 87 billion attacks from 19 million unique IPs by Vercel's WAF. 2. Enhanced Vercel Firewall with 40x faster DDoS protection through real-time stream processing. 3. Improved WAF responses by returning 403 Forbidden for persistent threats. 4. Introduction of Vercel Firewall DDoS mitigation notifications and enhanced Firewall data in the Monitoring tab. 5. Real-time audit log streaming for Enterprise customers to SIEM tools. 6. Completion of SAQ-D AOC for PCI DSS v4.0. The post also outlines future challenges like intensifying DDoS attacks, sophisticated bots, and the 'Human Layer' in security.
  12. Mitigating Denial of Wallet risks with VercelThis post details Vercel's defensive strategies against Denial of Wallet (DoW) attacks, which exploit cloud scalability to drive up operational costs. It explains how DoW attacks work by generating illegitimate requests below rate-limiting thresholds and exploiting auto-scaling applications. Vercel's defenses include granular API rate limiting, Vercel Web Application Firewall (WAF) with challenge-response protocols, real-time usage monitoring, budget alerts, spend limits, and advanced bot protection through behavioral analysis and challenge-based verification. The post emphasizes that these features protect both infrastructure and budget by filtering malicious traffic and ensuring scaling only for legitimate users.
  13. Life of a request: Securing your app's traffic with VercelThis post details Vercel's layered security approach for application traffic. It explains the platform-wide Vercel Firewall's multi-layered protection (L3, L4, L7) against DDoS, SYN floods, UDP floods, and HTTP floods. It also describes Deployment Protection for project-level access control (account, password, trusted IPs) and the customizable Web Application Firewall (WAF) with parameters like IP, geolocation, user agent, headers, and actions like logging, challenges, or blocking. It highlights the WAF's ability to enforce persistent actions to block repeat offenders and prevent unnecessary usage costs.
  14. Vercel WAF upgrade brings persistent actions, rate limiting, and API controlThis post introduces three new enhancements to Vercel's Web Application Firewall (WAF): persistent actions to block repeat offenders for a customizable duration, rate limiting to set request limits on specific endpoints and prevent API abuse, and a Firewall API for programmatic control of WAF settings. It also highlights the instant rule propagation (300ms) across the Edge Network and the extension of the Challenge page to the broader Vercel Firewall. Persistent actions allow edge requests to be processed earlier, bypassing usage metrics and WAF evaluation. Rate limiting can be configured per-endpoint and broadly, with layered checks. The Firewall API enables automation of security workflows and integration with third-party intelligence feeds.
  15. Preventing infrastructure abuse with Vercel FirewallThis post details Vercel's multi-layered security approach with the Vercel Firewall, highlighting its ability to block billions of suspicious TCP connections and HTTP requests weekly. It describes the mitigation of a 1.37 Tbps DDoS attack with no downtime, emphasizing the platform's adaptive defenses against evolving threats. The Vercel Firewall integrates Layer 4 and Layer 7 protections, leveraging application awareness from its framework-defined infrastructure for precise threat mitigation. It also introduces a new Firewall Rule template to block common AI bots.
  16. Protecting your app (and wallet) against malicious trafficThis post details how Vercel's Firewall can be configured with custom rules to block or challenge suspicious traffic based on IP addresses, JA4 fingerprints, and other criteria. It introduces Attack Challenge Mode as a temporary measure to investigate traffic spikes. The post also highlights real-time usage and spend controls, including setting spend limits and automatic app pausing. Code optimizations like enabling Fluid compute and transformation-based image optimization are recommended to reduce function and image processing costs. Moving large media files to Vercel Blob storage is suggested for cost-effective asset management. Finally, it reiterates the importance of default deployment protection for securing applications.
  17. Introducing the Vercel Web Application FirewallIntroduces the Vercel Web Application Firewall (WAF) as a new layer of defense against web threats. Key features include a customizable rules engine (path, user agent, IP, geolocation, JA4 fingerprints), framework-aware rules, managed rulesets for OWASP Top 10, observability into WAF metrics, rate limiting (beta), instant propagation (300ms), and instant rollback. The WAF leverages framework-defined infrastructure for intelligent edge request handling.
  18. Protecting AI apps from bots and bad actors with Vercel and KasadaThis post details the integration of Kasada's bot protection technology with Vercel's AI SDK Playground. It highlights the use of Next.js Middleware to enforce a 1st-party-request-only protocol, allowing for bot classification and blocking based on Kasada's analysis. The implementation resulted in a significant reduction of bot traffic from 84% to negligible levels. The post also discusses the threat vectors specific to AI applications, such as Denial of Wallet and Prompt Injection attacks, and Vercel's existing DDoS mitigation and Firewall capabilities.
  19. Introducing Vercel Firewall and Vercel Secure ComputeIntroduces Vercel Firewall with new custom firewall rules for IP blocking, enhancing DDoS mitigation and access control. Also introduces Vercel Secure Compute for Enterprise customers, enabling private network connections between Serverless Functions, builds, and backend infrastructure. Mentions seamless authentication through partnerships and developer-first tooling, and highlights Vercel Monitoring and Logs for continuous monitoring and audits.