About Strukto.AI
STRUKTO.AI is the structured layer for AI agents. We organize data, tools, and services into a structured substrate for AI agents, so they can actually do work, not just hold conversations.
Today's AI agents talk to dozens of services, run code in sandboxed environments, and orchestrate multi-step work over messy data. Each layer of that stack typically reinvents its own contract. Strukto gives agents one structured substrate underneath: a Lite Agent OS and a virtualization layer for AI agents. The same agent reads a Google Doc, queries Postgres, and ships a Linear issue with the same handful of primitives.
Who built Strukto?
Zecheng Zhang is the founder and CEO of Strukto and a researcher. He was a founding engineer at Kumo.ai, graduated from Stanford University and University of Illinois at Urbana-Champaign, previously engineered at Amazon Web Services (AWS), and is a Y Combinator alumni. His background spans ML systems, relational databases, structured data, Large Language Models, and AI agents.
Zecheng is a life-long open source contributor. He co-created DeepSNAP, PyTorch Frame, RelBench, and CRAB, and contributes to PyTorch Geometric, pyg-lib, CAMEL, and Eigent.
What does Strukto build?
| Layer | Product | Problem it solves |
|---|---|---|
| Environment | Mirage | A virtualization layer for AI agents that mounts S3, Drive, Slack, GitHub, Notion, Postgres, and Redis with the same Unix-like primitives |
| Agent System | AgentBox | A lite agent OS in the browser where agents do their work, with no server-side data storage |
| Evaluation | SEAR | LLM-as-judge evaluation and gateway routing as SQL over a typed relational schema |
MIRAGE
Available NowAgentBox
Available NowWhat is Strukto's research?
SEAR
PublishedHow do I contact Strukto?
Book a call: cal.com/strukto