RegVista MAS SG

Inspiration

I built RegVista because regulatory research is still surprisingly manual. You search for a document, open a few PDFs, work out which version is current, and then repeat the process when you need to compare requirements. I wanted to explore what changes when the regulatory website itself can be used by an AI agent.

What it does

RegVista MAS SG is a small regulatory-intelligence MVP focused on Singapore/MAS material relevant to digital payment token services. You can search the registry, open individual records, see their metadata and official MAS sources, and compare two records side by side. The same capabilities are exposed to AI agents through WebMCP.

How we built it

The app is intentionally lightweight: HTML, JavaScript and a small curated regulatory dataset, with three WebMCP tools:

• search_regulations
• get_regulation
• compare_regulations

The tools are registered directly in the browser using the WebMCP Imperative API.

The key workflow is simple:

search → identify the right records → retrieve or compare them

Challenges we ran into

The biggest challenge was working with an emerging browser standard while keeping the experience useful for a normal human user. I also wanted to avoid making the project look like an AI chatbot bolted onto a regulatory database. The agent should be able to use the same underlying capabilities as the human.

Accomplishments that we're proud of

I got the full flow working end to end: a live regulatory registry, a deployed public site, and three working WebMCP tools that an agent can discover and use. I am also particularly proud of the fact that the project remains simple enough to understand end to end while demonstrating a practical use case for the emerging agent-native web.

What we learned

WebMCP makes the boundary between a website and an AI agent more blurry. Instead of asking an agent to figure out how to navigate a UI, the site can tell the agent exactly what it can do. That feels particularly useful for structured domains like regulatory research.

What's next for RegVista MAS SG

This MVP is intentionally narrow. The natural next steps would be expanding the regulatory registry, supporting more jurisdictions and regulatory domains, improving document-level retrieval, and enabling richer agent-assisted research workflows. The broader idea is RegVista as an agent-accessible regulatory intelligence layer, with MAS Singapore as the starting point.

Built With

• ai
• chrome
• codex
• github
• html
• javascript
• regtech
• webmcp