Lyrie.ai exits stealth with $2 million to secure AI agents

Published: 2026-05-14T08:25:01.000Z

Lyrie.ai has come out of stealth with $2 million to build the security layer for AI agents. The Dubai-based platform is releasing an open cryptographic standard for agent identity and trust while preparing for a Series A targeting enterprise and government markets.

Every major enterprise deploying AI agents right now is doing so faster than any security framework has been able to keep up with. These agents are reading emails, writing and executing code, signing contracts, processing transactions, and operating across sensitive internal systems with broad access and, in most cases, limited oversight. The questions that should be answered before any of that happens, who is this agent, what is it authorized to do, and has it been tampered with, have largely gone unanswered because the infrastructure to answer them has not existed. Lyrie.ai , a Dubai-based autonomous cybersecurity startup developed by OTT Cybersecurity LLC, has come out of stealth with a $2 million pre-seed round and a direct attempt to fill that gap. The platform positions itself as the security infrastructure layer for the AI agent era, a framing that is specific and deliberate. Lyrie is not building another endpoint security tool or a conventional threat detection dashboard. It is building the trust and verification layer that sits underneath agentic AI systems, a category of infrastructure that did not need to exist three years ago and now needs to exist urgently. The $2 million raised from undisclosed investors will go toward expanding the security research team, scaling infrastructure, deepening enterprise and government partnerships, and advancing the submission of the company's centerpiece technical contribution through formal standards channels. That centerpiece is the Agent Trust Protocol, or ATP, an open cryptographic standard for AI agent identity, scope, attestation, delegation, and revocation. The protocol is royalty-free and is being submitted to the Internet Engineering Task Force, the body responsible for the technical standards that govern how the internet actually works. A reference implementation is available under the MIT license. Lyrie's founders are positioning the ATP as the trust layer for the agentic AI economy in the same way that SSL and TLS became the trust layer for the web, meaning a foundational standard that other systems build on rather than a proprietary product that locks in a single vendor. That open standards approach is a meaningful choice. It signals ambition at the infrastructure level rather than at the product level, and it creates a surface area for ecosystem adoption that a closed protocol would not. The platform's security capabilities span the full threat lifecycle. Automated penetration testing can execute a seven-phase test from a single command, generating proof-of-concept exploits and code-level remediation guidance. Adversarial AI red-teaming workflows run on H200 GPU infrastructure and support advanced AI security evaluation. Zero-day vulnerability research is handled through autonomous discovery workflows using binary analysis. The entire platform maps to the OWASP Agentic Security Initiative 2026 threat catalog, giving enterprise security teams a compliance and coverage reference they can communicate to boards and auditors. Alongside the funding announcement, Lyrie also confirmed that OTT Cybersecurity LLC has been accepted into Anthropic's Cyber Verification Programme, which supports verified cybersecurity operators working on advanced AI safety, red-teaming, and vulnerability research. The company is already preparing for a Series A aimed at scaling deployment across enterprise and government markets, which indicates the pre-seed is a foundation-building round rather than a runway extension. Lyrie is building toward institutional contracts, and the government angle is particularly relevant given how aggressively Gulf public sector entities are deploying AI tools across operations, services, and infrastructure. For the UAE and the broader MENA region, a Dubai-headquartered company building AI agent security infrastructure is well-positioned relative to where the market is heading. The UAE's AI adoption across government and enterprise is among the highest in the world by penetration rate, and the security risks that come with agentic deployment are landing here as fast as anywhere. Saudi Arabia's Vision 2030 digital transformation agenda and the UAE's AI strategy both create the kind of institutional demand that Lyrie is built to serve, and having that market on the doorstep from day one is a structural advantage that most cybersecurity startups building in this space from North America or Europe do not have.