Dubai’s Lyrie.ai Unveils $2M Pre-Seed to Secure Autonomous AI Agents

Mohammed Fathy

Last updated 13-05-2026

4 min

Dubai-based Lyrie.

ai raised $2m to secure autonomous AI agents.

It launched the “Agent Trust Protocol” to verify identity and permissions.

The royalty-free framework may be submitted as an open IETF standard.

Funds will back product growth, research hires, and enterprise expansion.

The firm also joined Anthropic’s Cyber Verification Programme, boosting credibility.

Dubai-based Lyrie.ai has stepped out of stealth mode with a $2 million pre-seed round, setting its sights on one of the biggest blind spots in today’s tech race: securing autonomous AI agents. The startup, developed by OTT Cybersecurity LLC, says the fresh capital will go towards product development, expanding its research team, and rolling out what it calls the Agent Trust Protocol (ATP) across its growing partner ecosystem.

For anyone building in AI right now, this definitely strikes a chord. I’ve lost count of how many founders across the MENA region tell us at Arageek that they are rushing to deploy AI agents into workflows, finance, logistics, even public services, without a clear framework for trust and verification. It’s exciting, yes, but also a bit of a faff when security is bolted on later rather than designed from day one.

Lyrie.ai wants to tackle exactly that. Headquartered in Dubai, the company focuses on securing autonomous AI agents that act inside enterprise systems, execute transactions, and access sensitive infrastructure. And believe it or not, much of this activity currently happens without robust identity verification for the agents themselves.

“The agentic AI economy is being built right now, and it is being built without a security foundation,” said Guy Sheetrit, CEO and Founder of OTT Cybersecurity LLC. He argued that many AI agents today operate without identity verification, scope enforcement, or tamper detection. According to him, Lyrie.ai has built infrastructure to address these gaps and structured it as an open standard for the wider industry to adopt.

That open standard is the Agent Trust Protocol. ATP is described as a royalty-free cryptographic framework designed to handle AI agent identity verification, authorisation, attestation, delegation, and revocation. In simple words, it aims to make sure AI agents can prove who they are, what they are allowed to do, and whether they have been altered. The company expects to submit the protocol to the Internet Engineering Task Force (IETF), positioning it as a potential open industry standard.

The ambition is bold. Lyrie.ai sees ATP as a trust layer for the emerging “agentic” AI economy, in much the same way SSL/TLS became the backbone of secure web browsing. That comparison is not small talk, SSL/TLS underpins every secure website connection today. On the flip side, building something that becomes a universal standard is no small feat. It requires adoption, consensus, and time.

Alongside the funding news, OTT Cybersecurity LLC has also been accepted into Anthropic’s Cyber Verification Programme (CVP). The programme backs cybersecurity operators focused on advanced AI safety, red teaming, and vulnerability research. It’s a notable endorsement in a field where credibility matters hugely.

As for the product itself, Lyrie.ai offers both offensive and defensive autonomous cybersecurity capabilities. This includes automated penetration testing, adversarial AI red-teaming, zero-day vulnerability research, and AI security evaluation tools. In other words, it aims to test and harden AI systems before attackers find the cracks.

The new funds will also support infrastructure scaling and deeper enterprise partnerships, with an eye on enterprise and government deployments. A Series A round is already being prepared to expand further into those markets.

From a regional perspective, it’s encouraging to see a UAE-based cybersecurity player lean into such a foundational layer of AI infrastructure. We often talk about flashy applications, chatbots, automation tools, you name it, but security is where the real long game is played. I reckon if AI agents are to become mainstream across enterprises and public institutions, trust frameworks like ATP will be not just helpful, but essential.

Whether Lyrie.ai’s protocol becomes the industry’s go-to standard remains to be seen. But stepping out of stealth with fresh capital and a clear thesis around agent identity is, at the very least, a move that feels spot on for this moment in AI’s rapid, and sometimes chaotic, evolution.

🚀 Got exciting news to share?

If you're a startup founder, VC, or PR agency with big updates—funding rounds, product launches 📢, or company milestones 🎉 — AraGeek English wants to hear from you!