Managing AI agents with Nyantrace.ai
Team: Poe Myint Swe (Tandon ’26), Niranjana Niranjana (Tandon ’26), Daniela Castano (Tandon ’26)
About the Venture: Nyantrace provides real-time monitoring and oversight for multi-agent AI systems to help prevent unintended or harmful behaviors. The platform tracks agent actions and interactions to improve safety, accountability, and operational control.
Tell us about Nyantrace and the problem it aims to solve.
We’re building Nyantrace, observability and governance for AI agents. Fully autonomous agents like OpenClaw that have agency and intent are going mainstream. Traditional observability tools are not prepared for agentic actions. We are going to fill that gap to get transparency behind every agent’s action, decision, and intent.
How did the team come together?
We’re NYU classmates who found each other through different group projects. We recognized each person’s unique skillset and decided to found a company. We experimented with product ideas like a personal assistant called Project Hermes and an autonomous alpha-trading platform called Project Argus. Nyantrace is our third idea from Project Aletheia, and the one we went all-in on. We believe guardrails and governance is the important segment of the agentic workflows and we want to be at the forefront of it.
What sets Nyantrace apart?
Most observability platforms today tell you something went wrong after the fact. We step in before it happens, catching what Agent is about to do and redirecting it if it conflicts with policy. Existing solutions watch each tool individually. We watch how they interact, which is where the real problems live. And with the EU requiring coordination oversight by August 2026, we’re already building the compliance trail most companies haven’t started thinking about.
What motivated you to apply to the Challenge?
We are NYU students. It’s only right that we launch while we’re in school. NYU has deep connections in the industries we’re going after, and there’s a community of founders tackling hard problems. The best learning happens when you’re surrounded by people building something real. We wanted to be part of that.
What has been the biggest turning point for you in your startup journey?
We started out doing industry research, user interviews, and building prototypes. The companies we talked to were just starting their agentic journey. No clear guidelines or policies were set. Then we realized what we actually needed to build: a system where policies are created from intent and precedent from overrides. We needed to capture the why behind every decision and let agents learn from it.
What have been the biggest challenges you’ve faced so far?
Getting our first company to say yes to a pilot. We’ve had great conversations and even some inbound interest, but at the enterprise level, it’s about relationships, not just demos. We’re getting closer. Once we have that first success story, the next ones will come much faster.
What are some recent milestones?
We’ve built the core product that plugs into a company’s agentic workflows, catches conflicts in real time, and gives teams a clear view of what’s happening—even if they’re not technical. That’s the piece we’re proud of. Right now, our focus is landing that first paying customer and proving it works in the real world.
What advice would you give to aspiring entrepreneurs, especially those just starting out?
Get comfortable hearing no. Most days, nothing works the way you planned. Most pitches don’t go anywhere. That’s normal. It’s easy to look at successful startups and think, “They have it all figured out.” But everyone starts with nothing. Everyone faces rejection. Everyone has days when they wonder if they’re wasting their time. The only difference between the startups that succeed and the ones that don’t is that the successful ones kept going.
