New platform, Trace-AI, moves beyond reactive scanning to predict and score threats before they are exploited, boosting vulnerability prioritization by 40% for early users.
LONDON, UK / ACCESS Newswire / October 6, 2025 / Zerberus.ai, an innovator in AI-powered cybersecurity, today introduced Trace-AI, the industry’s first predictive supply chain security platform. It helps organizations proactively find and neutralize threats by understanding not just what is in their software, but how it could be attacked.
Software development is at a turning point. The recent XZ Utils backdoor showed that today’s security tools are blind to modern threats. They can find known CVEs but miss sophisticated attacks like social engineering or malicious project takeovers. These tools create long lists of components but can’t spot the subtle signs of an attack, leaving teams struggling to know which packages are a real, immediate threat.
“For too long, security has been about chasing alerts after the damage is done. The future is about preventing breaches before they happen,” said Ramkumar Sundarakalatharan, Co-founder of Zerberus Technologies. “Trace-AI is built on serious academic research. It gives companies a predictive advantage, helping them see and act on the hidden risks in their open-source code.”
Introducing Predictive Supply Chain Security
At its core, Trace-AI uses ZSBOM, an open-source framework developed from university research. This allows Trace-AI to analyze many different risk factors beyond just CVEs and give every component a single, clear score.
Key features include:
- Maintainer Anomaly Detection: Flags suspicious project takeovers by analyzing maintainer profiles and contribution patterns. It learns from real-world incidents like the XZ Utils backdoor to spot threats that normal scanners can’t see.
- Proactive Threat Detection: Finds new threats early, including typosquatting attacks that trick developers into using fake packages like python-urlibs instead of real ones.
- Multi-Dimensional Risk Scoring: Replaces noisy alert lists with a single risk score (0-100). The score considers a component’s potential for attack, its stability, the maintainer’s reputation, and community trust.
- Complete Transitive Visibility: Maps your entire dependency tree, including dependencies of dependencies, to find risks buried deep within your software.
Sriram G, Head of Product at Zerberus, comes from the high-stakes world of fintech. There, he led engineering teams where top code quality was essential for survival.
“Engineers are judged on the quality and speed of their work. Security tools have historically slowed them down, flooding pull requests with alerts that lack context,” he said. “We are changing that. Our view is simple: a dependency risk is a code quality issue. We designed Trace-AI as an engineering tool first. It works inside the CI/CD pipeline to give a clear, reliable signal, not more noise. It tells you ‘this dependency is a real, exploitable risk to your app,’ so you can fix what matters and merge your code with confidence.”
Early adopters are already reporting great results. Over 35 startups and 150 developers are using Trace-AI, with users seeing a 40% improvement in prioritizing vulnerabilities and 35% faster incident response times.
Availability & Roadmap
Trace-AI will be publicly available for direct signup starting Wednesday, 8 October 2025. Teams can create an account at https://zerberus.ai
The platform launches with full support for Python, Node.js (npm), and Java. Support for Go, Scala, and Rust is planned for a pre-release by Christmas, with more languages on the roadmap for 2026.
About Zerberus Technologies
Founded in 2024, Zerberus.ai is a pioneer in predictive supply chain security. Its products, including Trace-AI™, Compl-AI™, and Remed-AI™, help companies move from reactive to proactive security. The company is extending its expertise into the AI era with its upcoming platform, RAGuard-AI. With operations in the UK and India, Zerberus is trusted by over 70 enterprises and SaaS builders.
Media Contact:
Zerberus Technologies Ltd
Ramkumar Sundarakalatharan
