About Logic Pulse Lab
Logic Pulse Lab is an AI-native product lab focused on Agent systems, IM interactions, open-source tools, and experimental products.
Agents should enter real workflows.
We believe AI Agents should not stay only in chat windows. They should understand context, call tools, proactively follow up, and help users complete long-term tasks inside real communication and workflow surfaces.
The goal of the lab is not to package concepts, but to validate AI Agent capabilities through real products, real users, open-source systems, and shipped product experiments.
Li Zihao / YDYK
YDYK is the founder, builder, and research-product lead behind Logic Pulse Lab.
Li Zihao / YDYK focuses on full-stack product development, AI Agent systems, multi-channel IM gateways, automation tools, and open-source products. Through the lab, he continues to build Coke / OpenCoke, input enhancement tools, MCP tools, and selected customer AI systems.
A focused product and systems surface.
AI Agent systems
IM gateways and conversational interfaces
Reminder and follow-up systems
Tool calling and skill loading
Memory and context systems
Open-source developer tools
Product experiments
Selected client systems
A small lab with a wider delivery network.
Logic Pulse Lab is led by YDYK, and collaborates with specialists when project scope requires frontend, backend, design, or deployment depth.
Founder, full-stack builder
Focuses on AI applications, agent systems, C++ systems work, and full-stack product development.
Li - former Microsoft frontend engineer
Frontend specialist
Works on modern frontend frameworks, engineering quality, and high-performance user interfaces.
Wang - former Tencent backend engineer
Backend architect
Supports backend architecture, AI deployment, database design, and scalable service systems.
Zhong - former NetEase designer
UI/UX designer
Contributes product interface design, interaction design, prototyping, and user experience work.
Some original capability signals are still part of the story.
These are conservative signals from the previous site, kept as delivery context instead of sales claims.
shipped projects
collaboration clients
years of building
How the lab chooses work.
Build real products, not only demos.
Ship small and verify in real usage.
Prefer open, inspectable, and reusable systems.
Keep AI close to everyday workflows.
Turn experiments into tools and systems.