Cloudflare launches Precursor to stop evasive bots
Cloudflare, Inc. introduced Precursor, a behavioral validation engine that monitors entire user sessions to detect and block advanced bots without disrupting legitimate traffic. The tool addresses the fact that bot traffic now accounts for 57% of all web requests.

*this image is generated using AI for illustrative purposes only.
Cloudflare, Inc. has announced the general availability of Precursor, a next-generation, continuous behavioral validation engine designed to stop evasive bots without disrupting legitimate users. Built directly on Cloudflare’s edge, Precursor runs inside web browsers to monitor entire user sessions, detecting bot automation by analyzing ongoing interactions in real time rather than relying on static checkpoints. This development addresses the shift in internet traffic, where automated bot activity has eclipsed human activity to generate roughly 57% of all web requests.
Continuous Behavioral Validation
Traditional security checks often evaluate a single moment in time, which modern bots can bypass by faking a single action. Precursor addresses this by analyzing telemetry across an entire session to unmask automated imposters. The system provides a session-level view of site activity by collecting robust browser signals to block unwanted automated traffic. This approach eliminates the blind spots that exist between critical security moments like login and checkout.
Key Features
Precursor offers several distinct capabilities designed to enhance security while maintaining user experience:
| Feature | Description |
|---|---|
| Privacy-Led Defense | Logs aggregate behavioral patterns rather than specific user inputs. Keyboard activity is recorded as timing rhythm, never capturing actual keystrokes. |
| Zero-Code, One-Click Setup | Enabled with one click, injecting a dynamic script without modifying underlying code. Evaluates mouse movement, scrolling rhythm, typing cadence, and clipboard activity. |
| Real-Time Analysis Engine | Servers instantly unpack telemetry data to scan for faked activity, validating interaction streams against human behavior patterns. |
| Session-Long Security | Continuously evaluates the user journey across web or single page applications, preventing bots from resetting signatures by refreshing pages. |
Operational Context
The rise of AI agents has transformed the internet from a landscape built for human clicks to one dominated by automated threats. These threats drive up infrastructure costs, manipulate inventory, and compromise data. By moving away from static, point-in-time defenses to continuous behavioral validation, organizations can better protect their integrity against this new breed of automated threats.
How will the introduction of continuous behavioral validation impact Cloudflare's competitive positioning against other edge security providers?
What are the potential implications for legitimate AI agents and automated tools as Precursor becomes more widely adopted?
Could the computational overhead of real-time telemetry analysis affect Cloudflare's edge latency or pricing models?
































