Contact

Swarm Intelligence for Cyber Security

Continuous and affordable attacker‑grade security.

AI changed what it costs to attack. We change what it costs to defend. swarm.security builds AI agents that test the websites and APIs your company runs, all year, the way an attacker would.

Product one

Scheduled pentesting

Our agents pentest your websites and APIs on a schedule you set. They work from the internet, as an attacker would, so we never need your source code. A security researcher reviews every finding before you see it.

01

You say what to test

You choose the websites and APIs in scope and how often they're tested.

02

Agents test from the outside

Each finding comes with a replayable exploit chain, so your team can rerun every step.

03

Reviewed findings, then a retest

You get a log of every request we made, and we retest your fix on the next run.

What comes next

One company for the whole of security

Scheduled pentesting is our first product. We're building toward detection, response and purple teaming, where our agents attack and help you defend at the same time.

Founders

Who we are

swarm.security has two founders. We're building the company in the Antler Japan residency in Tokyo.

Michalis Vasileiadis

Co-founder, research and product

Michalis is the researcher behind our public CVEs, including bugs in Samba and OpenSSH. He reviews every finding before a customer sees it.

Rayyan Zahid

Co-founder, sales and go-to-market

Ray has built agent orchestration since before LLMs. He is a Nebius Fellow and founded Immersive Commons in San Francisco.

Contact

Talk to us

If your company has to pentest, or you want to try scheduled pentesting, write to us.

ray@swarm.security