COMPANY

Built for the
agentic era.

Security tools were designed for humans running manual workflows. We're building for the next era, where autonomous agents and CI pipelines need machine-readable answers, not PDF reports.

MISSION

Security should be autonomous,
not aspirational.

Diverg was founded on a simple observation: security investigations are still mostly manual. A team discovers a vulnerability, files a ticket, and waits weeks for a fix, while the window for exploitation stays open.

We bridge web security and blockchain forensics into one platform that operates autonomously. Whether you're assessing a target before a deal, monitoring production for regressions, or tracing funds after an incident, Diverg delivers evidence-backed findings without human babysitting.

01
Fact-only output
We never surface placeholder data as findings. Every result is verified or clearly marked as skipped.
02
Agent-first design
Machine-readable JSON, structured severity, and CLI gates that integrate into any pipeline without glue code.
03
Open by default
Diverg Open is MIT licensed. Audit the probes, extend the rules, run it on your own infrastructure.
3
Investigation surfaces
6
Active probe modules
100%
Fact-only findings
MIT
Open source

VALUES

What we stand for.

Evidence, not vibes

Every finding ships with proof: the payload that triggered it, the response that confirmed it, and a specific remediation, not generic advice.

Multi-surface by default

Web security and blockchain forensics are separate disciplines solved by separate tools. We think that's wrong. Attackers don't stay in one lane.

Speed without shortcuts

Active probing in under 30 seconds. We achieve this with parallel execution and stealth-aware rate limiting, not by skipping checks.

Open and auditable

Security tooling that can't be audited shouldn't be trusted for security. Our core scanner is public, documented, and tested.

Join us in building it.

We're hiring across engineering, design, and marketing.

See open roles Get in touch →