Skip to content
Companies

Runsybil

AI agent that runs continuous penetration tests

Raised to date

$30M

of $30M offered across 1 Form D filing

Still open
$21
Investors
5
Last filed
Apr 3, 2026

About

RunSybil builds an AI-native offensive security platform that continuously pentests live applications and infrastructure. Its AI agent, Sybil, autonomously discovers, exploits, and documents real vulnerabilities by interacting with systems through their external interfaces rather than scanning source code. The company was founded in 2023 by CEO Ari Herbert-Voss, OpenAI's first security research hire, and CTO Vlad Ionescu, formerly head of offensive security red teaming at Meta. It raised $40M led by Khosla Ventures in March 2026, with S32, the Anthology Fund, Conviction, and Elad Gil participating.

Adversarial reasoning rather than signature-based scanning: the agent chains weaknesses across code, APIs, and cloud infrastructure into validated, exploitable attack paths, and runs on every deployment instead of on an annual pentest cycle, so teams get only pre-validated findings.

Funding

  • EquityFiled Apr 3, 2026Form D
    $30M/ $30M

Product

Sybil

Offensive security / automated penetration testing platform

An autonomous AI agent that performs continuous penetration testing against live applications, APIs, and cloud infrastructure, exploiting and documenting real vulnerabilities without a human in the loop.

  • Autonomous exploitation of live systems through their external interfaces, not static source analysis
  • Full-stack mapping across code, APIs, and cloud infrastructure to catch integration-point vulnerabilities
  • Security feedback on every pull request instead of periodic rescans

Built for

Engineering and security teams at high-growth technology companies, financial institutions, and Fortune 500 enterprises running high-risk applications; named customers include Notion, Cursor, Carta, Baseten, Turbopuffer, and Thinking Machines Lab.

People on file