Thesis

We started from a simple belief: retail is undergoing another major shift. After the move from brick-and-mortar to e-commerce, the next transformation is toward 3D imaging and virtual fitting.

From my experience managing the tech aspects of a jewelry e-commerce business, I saw firsthand how much better engagement and conversion rates became when we transitioned from 2D photos to animated 3D visuals. But building those assets was expensive and complex. Only large brands like Rolex could afford to pay 3D studios hundreds of thousands of dollars for custom models. Mid-tier e-commerce businesses didn’t stand a chance.

We wanted to change that. Our vision was to create a no-code, open-source platform that would let anyone integrate sophisticated 3D models into their websites easily. Using the JavaScript library Three.js (and potentially wrappers like Didai), we aimed to give small and medium-sized businesses access to advanced rendering capabilities without massive costs or infrastructure.

At first, we imagined a visually striking product: shimmering effects, rich texturing, and an AI-first layer for generating animations. We planned to keep the base platform open source, while monetizing through a SaaS layer that hosted and optimized the heavy rendering work. Customers would pay for managed infrastructure instead of building it themselves.

Because of my background, we decided to start with jewelry, a space where we already had expertise, relationships, and proven demand. The idea was to build credibility and traction in a niche market before expanding to broader categories like furniture or fashion.

Validation

We began validation through direct conversations – about a dozen in total – with shop owners, competitors, and professionals from jewelry and e-commerce companies. My co-founder led most of this outreach.

We intentionally focused on the U.S. market, where the opportunity and scalability were greater. From early discussions, we learned that working with Israeli jewelers would be difficult; their expectations for visual perfection were extremely high. Starting with the U.S. gave us a better chance to iterate quickly.

We built and presented a working demo, which sparked real interest among potential users. The demo included:

1. 3D Sophistication: Rendering advanced 3D visuals beyond basic 360° views, with shimmering and realistic material texturing.
2. Cost Efficiency: A shared catalog of material layers (like leather or gold textures), designed so customers wouldn’t need to pay separately for each asset.
3. AI Integration: Plans to use AI for automated animations, such as rotating models or moving them in and out of frames.

However, during technical testing, we ran into a major obstacle: the tech simply wasn’t “there yet.” Jewelry clients demand perfect, ultra-detailed images, and our solution couldn’t yet meet that standard at a reasonable cost. We tried pivoting toward furniture, a niche with looser visual requirements, but even then, the technical and financial barriers remained.

We were bootstrapped from day one and deliberately avoided fundraising at first. We spoke with two investors but made it clear we weren’t raising yet, we wanted user traction first. Despite positive feedback on the demos, the product never reached a fully sellable state. After about six months, we decided to pause.

Why it didn’t progress

1. Technical Limitations and Cost: The technology turned out to be far more complex and expensive than we expected. Even after months of work, our rendering quality couldn’t meet the jewelry industry’s standard for flawless detail. We pivoted to furniture, hoping for easier technical requirements, but the challenges persisted.

2. AI Leap, Competition, and Timing: We began to see that emerging AI models might undercut our value proposition entirely. Larger players, for example, R&D teams at AI or e-commerce startups like Triple Whale could easily develop similar animation tools faster. We also realized that many retailers might prefer interactive videos (AI-generated from 2D images) over full 3D models, reducing the immediate demand for what we were building.

3. The Bootstrapping Limit: At some point, it became clear that we couldn’t bootstrap this. To continue, we would have needed to raise significant capital to build a highly skilled dev team, a risky move given the uncertain demand. 

Insights

1. Integrate Business and Technical Roles: We learned that the business side must deeply understand what’s happening technically. In early-stage startups, silos kill learning. It’s crucial for founders to cross over, at least a little, and understand each other’s domains, during these early stages.

2. Community Is the New Moat: In today’s AI-driven world, code isn’t the strongest advantage, but community is. Building an open-source infrastructure with a strong community creates real defensibility and sustained value.

3. Use Advisors Early: Even when bootstrapping, I now believe it’s worth bringing in an angel investor early, not for the money, but for introductions and guidance. Advisory shares can be a powerful way to surround the team with people who open doors.

4. Validate Technical Hurdles First: If I could go back, I’d start by identifying the hardest and most expensive technical problems before anything else. It would have saved months of effort and made our decision points clearer.

5. Timing and Foundational Tech Risk: The 3D space will matter, especially as virtual and immersive retail grows, but it still needs time to mature economically. Building in such a space means betting on a future theory. And if general-purpose AI models start replicating your core functionality, you need a much stronger differentiator.

Resources:

Download the business plan

---