We are now raising a $3M pre-seed round from early-stage VCs, angels, and the top accelerators this fall.
Vulnzap, our first product, is in private beta with public launch set for 1 August 2025. Early momentum:
more than 650 developers have installed the open‑core scanner for private testing
weekly growth is about twenty‑two percent
over 2.3 million lines of code have been scanned and seven thousand critical issues auto‑fixed in real time
five enterprises are running private pilots, two include paid proofs of concept, with signed contracts queued for GA
We begin with code, yet we are building for something far bigger, something no one has dared to make truly possible.
In 2024 we tried to build agentic hardware straight out of high school with zero capital and zero permission. We launched in the Emirates, watched the first team shatter, and absorbed every hard lesson at top speed. What rose from those ashes is sharper, tighter, and pointed directly at real problems. Scar tissue became focus.
The Vision
The future will not belong to whoever has the biggest cluster or the loudest headline. It will belong to the builders who own their intelligence completely.
Most AI today is a rented mind. Log in, accept the black box, and leave your ideas behind someone else's glass. Plaw exists for the opposite. We deliver a foundation of sovereign intelligence that any team can hold, shape, and deploy without compromise, permission, or hidden doors.
Sovereign intelligence in practice means you can:
run the model on your own hardware or cloud
audit every layer of the stack, from weights to logs
retrain or replace any component at will
unplug at a moment's notice and keep full functionality offline
Your stack. Your rules. Permanently.
Why Now
AI velocity is ripping past traditional security and policy cycles. Vibecoding—shipping software by prompting large language models at machine speed—adds millions of new lines overnight and multiplies hidden attack surfaces.
Vulnzap answers that urgency: a real‑time immune system that lives in the IDE, partnering with every keystroke instead of scanning after the fact.
This is Chapter One. Chapter Two is the broader sovereign‑AI stack: self‑hosted models, agent orchestration, and provable safety layers that keep the builder—not a cloud vendor—in the driver's seat.
We are not a lab. We do not sell research papers or vapor. We ship tools for people whose work cannot wait.
The Raise
Round: $3M pre‑seed (SAFE, $12M post‑money cap)
Runway: 18 months
Use of funds:
55 percent hiring (3 senior engineers, 1 product designer, 1 enterprise AE)
20 percent go‑to‑market and community (developer relations, vibecoder advocacy, SF headquarters setup)
15 percent compute and security audits
10 percent contingency and legal
Milestones on this runway:
GA release of Vulnzap (Q4 2025)
$1M ARR run rate (Q2 2026)
MVP of self‑hosted sovereign‑model stack (Q2 2026)
10 enterprise customers in production (Q3 2026)
If you back technical founders tackling foundational AI risk, let us show you the plan.
Team
Yaz Celebi (Co‑Founder & CEO) – product, design, relentless generalism
Anirudh Patel (Co‑Founder & CTO) – systems, LLM infrastructure, security engineering
Sherif Elamir (Founding Engineer) – full‑stack velocity from kernel to UI
Kyrie Kirk (Head of Marketing & Sales) – turning interest into signed contracts
Fully remote across Turkey, India, Egypt, and Canada; incorporating in Delaware and relocating to San Francisco.
Join Us
We are hiring senior backend engineers, an ML alignment specialist, and a product designer who want to put new powers safely in human hands. Bring your edge and your skepticism—we will use both.