TrustAI Laboratory
TrustAI aims to be a guardian of the GenAI era by reducing AI misuse and fraud through prompt defenses and content-filtering technologies. Its platform evaluates AI applications for security and safety risks, protects them with real-time controls, and provides red-teaming capabilities.
At a Glance
- Enterprise AI application providers
- Security, risk, compliance and AI engineering teams
- LLM and GenAI developers
- AI researchers and red teams
- +1 more
AI Tools by TrustAI Laboratory
(1)
Learn Prompt Hacking
LLM Prompt Security Learning Repo
Discussions
No discussions yet
Be the first to start a discussion about TrustAI Laboratory
Latest News
TrustAI hosted GenAI for SEA - 2025 Outlook in Singapore.
Learn Prompt Hacking repository updated with agent-evaluation dataset material.
TrustAI updated its Website Prompt Injection Demo repository.
TrustAI updated its OneRouter repository, described as a unified API for hundreds of AI models with fallbacks and cost-aware selection.
Products & Services
Model-agnostic evaluation for LLMs and GenAI agents, using blended prompt-injection fuzzing and other attack methods, with detailed reporting and recommendations.
Real-time protection for GenAI application providers that identifies the intent behind user inputs and AI outputs, with low-latency controls, multilingual support and customizable policies.
AI red-teaming service using adversarial prompt-fuzzing techniques and a large attack-data corpus; the company also provides introductory AI red-team materials.
An LLM vulnerability scanner and zero-day vulnerability fuzzer with a desktop GUI, manual and automated red teaming, evaluation datasets, reporting and a jailbreak cloud service/API.
Market Position
TrustAI positions itself as a Singapore-based, end-to-end AI application security provider combining model-agnostic evaluation, real-time protection, adversarial red teaming and open educational resources. Tracxn identifies Enkrypt AI, Protect AI and Virtue AI as top competitors; TrustAI differentiates through its prompt-fuzzing/red-team emphasis, community CTF and course materials, and LMAP's desktop-plus-API workflow.
Leadership
Founders
Lawrence Xu
Former CTO of StarCross and senior researcher at Alibaba; led $40M on-premise AR and $60M SaaS ARR products and published in BlackHat, DEF CON, HITB, BlueHat, XCon, KCon, IJCAI and RAID.
Andrew Hann Zheng
Former principal engineer at Alibaba with 10+ years in AI R&D; has participated on committees for global data-mining and CTF competitions and published in DataCon, BlackHat, ICSE and XCon.
Andy Du
Former co-founder and COO of StarCross; Ph.D. in computer science from the University of Missouri and MBA from Oxford; directed an NSA-funded project, published at ACM CCS, and worked on a G42 partnership.
Executive Team
Lawrence Xu
Co-founder and CEO
Former CTO of StarCross and senior researcher at Alibaba; led large on-premise and SaaS products and has extensive security-conference publications.
Andrew Hann Zheng
Co-founder and CTO
Former Alibaba principal engineer with 10+ years in AI R&D and publications in DataCon, BlackHat, ICSE and XCon.
Founding Story
TrustAI was started in response to the rapid adoption of AI applications and the resulting security and safety risks. Its initial vision was to align AI for a trusted future by combining evaluation, protection, red teaming and community education.
Business Model
Revenue Model
Commercial AI-security software and services: platform-based evaluation and protection, AI red-teaming services, and API/SaaS access to LLM testing and jailbreak capabilities. The company advertises a free start and demo-led access.
Target Markets
- Enterprise AI application providers
- Security, risk, compliance and AI engineering teams
- LLM and GenAI developers
- AI researchers and red teams
- Organizations deploying agents and custom or open-source models
- Evaluating the security and safety of enterprise LLM applications and agents
- Real-time filtering and protection for GenAI application providers
- Manual and automated AI red teaming and vulnerability discovery
- Continuous evaluation, regression testing and CI/CD security testing
- Training developers, researchers and security teams in prompt hacking and LLM security