I build systems that make complex things easier to change.

Asa Tsai · Software architect · Taipei

10+ years building frontend systems, design tools, and the architecture behind them.

I keep coming back to one question: why should adding one feature require understanding the whole application?

Years before Asyra, I proposed this direction at my previous company. It wasn't adopted, but the idea stayed with me. After leaving, I wrote it down as Communication-Driven Development (CDD): start with what needs to be communicated, and make responsibilities clear to both the system and the people building it.

Asyra grew out of those principles. I'm building it so the next person can understand, change, and extend a product without needing me beside them.

Read the CDD docs · Docs on GitHub

One foundation. Different worlds.

Asyra handles the shared infrastructure: editable state, transactions, Undo/Redo, and clear ownership. Each app brings its own rules and its own world.

Design

A canvas editor with editable vectors, Undo/Redo, and optional AI and collaboration. A working product built on the same foundation.

Try the editor · Source · Follow the architecture

Simulation

I'm starting with a robot workcell to build the architecture for a larger goal: simulating factory processes, including semiconductor fabs, to find problems before they happen on the floor. Over time, I want to bring my earthquake engineering background into building and factory safety analysis too. The current workbench explores geometry, clearance, and motion; it's an early development checkpoint.

Try the workbench · Source

FieldScope

A greenhouse modeling workspace with structures, crops, and a robot concept. The scene exists; patrol and harvesting simulation are still ahead.

Explore the greenhouse · Source

Make it easier to pick up the work.

The same question shows up in my other tools:

• handoff-ai keeps project knowledge available across people, agents, and sessions.
• context-rag finds the project context an agent needs for the task at hand.
• local-leetcode-trainer brings coding practice and AI guidance into a local workflow.

I care about clear boundaries, predictable changes, and leaving enough context for whoever comes next. A good demo gets my attention. Being able to keep changing it is what makes me stay.

---

Taipei, Taiwan · Remote collaboration

Let's talk on LinkedIn · Read the Asyra docs