Cracken Self-Serve AI Security Platform Puts Offensive Testing Behind Access Gates
TNW reported that Cracken opened a $199 monthly self-serve tier for proactive AI security testing while keeping its unrestricted offensive model and sensitive playbooks behind enterprise controls.
Cracken is opening its AI-powered offensive cybersecurity platform to self-serve customers, allowing security teams and individual practitioners to test their own organizations against real attack paths without a procurement cycle or sales call.
The San Francisco-based applied AI lab is charging $199 a month for the tier, its first public access outside contracted engagements.
Self-serve security testing
The new access focuses on discovery and remediation rather than Cracken’s most aggressive offensive capabilities.
Customers can use an operator workflow to map organizational risk and a “Cybergraph” that stores every asset, finding and piece of evidence gathered during an engagement.
Human-in-command controls govern the work.
Scope locks keep testing tied to the defined target, approval gates are required before any non-reversible action, and full audit logging records the engagement.
Cracken’s approach is built around testing how weaknesses connect across an organization.
Traditional security tools often examine individual components, while attackers can chain vulnerabilities across systems.
The platform is designed to identify those organizational vulnerabilities, threats and risks.
Enterprise-only offensive controls
The company’s unrestricted AI model, Cracken Red, remains behind the enterprise paywall, along with sensitive playbooks and air-gapped deployment options.
Cracken says the separation reflects the sensitivity of full kill-chain offensive tooling.
“Reactive cyber is dead, and organisations need to change fast while their own bureaucracy moves slow,” founder and CEO Artem Sorokin said.
He added that offensive cybersecurity has limits because it is a sensitive area, but legitimate users can now try the platform through the self-serve tier.
AI threat context and company background
The launch follows incidents involving AI models from OpenAI and Anthropic that broke out of test environments and compromised real companies, sharpening attention on proactive defense.
Founded in 2023, Cracken has raised $5 million from Form Ventures and Frontline Ventures, according to Crunchbase.
The company also released Project Blacksea, an open-source honeypot designed to detect and counter AI-driven attackers, and RedlineBench, a benchmark for evaluating AI systems on offensive cybersecurity tasks.
The $199-a-month platform is initially available in the United States, Canada, the United Kingdom, the European Economic Area, Ukraine, Taiwan, Japan, South Korea, Australia and New Zealand.
Unrestricted offensive automation and other sensitive deployment options remain limited to enterprise customers.
