When Institutions Enter Your Passion Project Space
The data.zeeker.sg series
This is Part 1 of a 3-part series on building and potentially closing down data.zeeker.sg:
- Part 1: When Institutions Enter Your Space (this post) - The decision framework and where I am now
- Part 2: Technical Lessons - The full technical stack, what worked, what didn't, and how AI tools made solo infrastructure building feasible—for builders who want the implementation details
- Part 3: What I Learned at SMU + My Decision (coming after Nov 18) - What they're building, how I decided, and what happened next
Ever since I was a law student, I dreamt of a world where judgments were data, and that it was accessible to everyone including poor students like me. I've grown older now, but that dream has taken me through the counters of AI and technology.
data.zeeker.sg is my latest attempt. I built it. It works.
Then SMU Centre for Digital Law announced they're launching a legal database project. Institutional resources doing what I tried to do solo.
The pragmatic voice in my head says: Stop. Institutions can do this better.
But I'm still in love with this vision.
What This Project Means to Me
This isn't just about data.zeeker.sg. It's about a full-stack vision for legal information infrastructure.
The idea: start from structured data and build whatever apps you can imagine on top. A daily aggregator that becomes a weekly podcast. An end-of-year review generator. A chatbot. An MCP server for Claude Code integration. More data sources—not just legal news, but judgments, legislation, regulatory notices. Eventually: AI-powered legal services. Maybe even rethink what a law firm could be if it's built on this kind of infrastructure.
That's the dream. Not just making legal information accessible, but creating the foundation that lets anyone build tools on top of it.
Why I Built This
Singapore's legal information exists, but it's scattered. Say you read a legal news article about a recent employment case. You want to read the actual judgment—that's in LawNet or a PDF somewhere. You want to check the statute it interpreted—that's on a different government site. You want to see related commentary—that's scattered across a wide range of different legal blogs and news sources. For developers trying to build tools that connect these dots, it's a nightmare of different formats and scraping challenges. You can browse each source manually, but you can't build integrated tools that connect the pieces.
I wanted to see if one person could create infrastructure that used to require teams. GenAI made it feasible—Claude helped me build it all.
The technical reality: I built Python scrapers that check an RSS feed daily, store articles in a searchable database with full-text search, and serve them via a REST API. I collected 346 legal articles and proved the stack works. The infrastructure runs: automatic updates, structured JSON data, machine-readable format.
I proved it's possible. One person can build this now.
Houfu Ang
Then SMU Announced Their Project
SMU Centre for Digital Law is launching their legal database project. Hiring full-time staff for 3 years. "Singapore's first open empirical legal database."
I don't know what they're actually building yet. Their scope, their approach, their technical infrastructure—all unknown until their November 18 event.
But institutional resources doing what I tried to do solo changes the equation. It's not a competition, and I do think they have a better shot at succeeding where I failed.
The Solo Builder's Dilemma
I'll admit that I never announced data.zeeker.sg. I kept thinking "needs more resources before it's ready." I built the infrastructure but waited for it to feel "complete enough" to share publicly. Zero known users because I never really tried to find them.
Now renewal forces the question: If I'm not willing to announce it, why am I maintaining it? Now that the competition has moved in, is this vision still mine to pursue?
But let's not forget the immense challenges I face. I still have to mind my day job and other commitments. In spite of that, I managed to eke out over 150 hours across several months to build this infrastructure. Even if I can accept that I managed to make something, this was an immense and costly undertaking.
The Two Voices
Two voices keep arguing in my head:
Pragmatic voice: Over 150 hours with zero known users. Never validated demand. SMU has institutional backing and dedicated staff. You proved it's technically feasible—that's enough. Time to let it go.
Passion voice: I'm still in love with this vision. This is labour of love with different success criteria. Zero users might mean poor distribution, not bad vision. Institutions existing doesn't mean quit. I could pivot to apps while they handle infrastructure, or do complementary work.
I don't know which voice is right. That's why I need a framework.
A Framework for Deciding: Three Questions
When institutions enter your passion project space, three questions help cut through the noise.
1. Assess the Overlap
I need to understand what SMU is actually building. Until I attend their November 18 event, their scope, approach, and technical infrastructure remain unknown.
Key questions:
- What's their scope? Legal news like mine, or case law and empirical research?
- Who's their audience—researchers, developers, practitioners?
- What's their technical approach—APIs, bulk downloads, just a website?
Until I know their scope, I can't know if we're building the same thing or something adjacent.
2. Identify Remaining Gaps
Even with overlap, gaps might remain. Maybe they're comprehensive and I can be experimental. Maybe they target researchers and I target developers. Maybe they handle data infrastructure and I focus on the application layer. Or maybe there's complete overlap and I gracefully hand off.
3. Clarify Your Motivation
This is the hardest question. What am I actually trying to achieve?
The Motivation Question
If this were a business, the answer would be clear. Zero users means stop or pivot. Institutions with more resources mean competitive threat. Find product-market fit or shut down.
Houfu Ang
But this is labour of love. My day job is for income. I build this because I want to see if it's possible, because I think Singapore needs it, because I'm still in love with the vision.
When I'm honest with myself: I don't know if anyone wants structured legal data APIs for Singapore. Am I building this because:
- Singapore needs it? If SMU provides it, maybe my work is done. Mission accomplished, even if it wasn't my infrastructure that succeeded.
- I need it for my practice? Then building for myself is enough. User count doesn't matter.
- I want to prove I can? Already proved it. One person can build this now.
- I'm still exploring what's possible? Then continue makes sense, regardless of users or institutions.
I think it's the last one. I'm exploring. But I'm not sure. I don't have a good answer yet.
I Don't Know Yet
I'm attending the SMU Centre for Digital Law event on November 18. After that, I'll have information to actually decide.
The domain renewal isn't until December. I have time. Server costs and all? I still have the resources to try.
What I know now: The question isn't about $110. It's about what I'm trying to achieve with this vision I've carried for years, and whether I'm building for the right reasons.
And if I do stop? There are still many options I can consider: Open-source the code so others can learn from it. Offer the data pipeline to SMU if they want it.
Whether I continue or stop, the technical lessons from building this solo are worth sharing—what worked, what didn't, and how AI tools changed what's possible for single builders. And whatever I learn at SMU's event might help other solo builders navigate similar crossroads when institutions enter their space.
Shutting down doesn't have to mean failure. It can also mean intentional closure of one chapter before starting the next. For other solo builders facing similar moments: when institutions and businesses enter your space, you'll face the same voices. Pragmatism says cut losses. Passion says keep going. The decision depends on what gaps remain and what you're actually trying to achieve.
Member discussion