From frustration to open source adoption: my journey building SnapDOM

3 points

8 months ago

I started hacking on what later became SnapDOM in April this year. At first it wasn’t even meant to be a standalone library — it began as an internal tool for a zoomable UI project (Zumly), I was building. To make that project work, I needed a way to capture DOM elements with high speed and fidelity. Existing DOM-to-image libraries like html2canvas often failed in subtle but important ways: missing pseudo-elements, mis-rendering fonts, ignoring Shadow DOM, or breaking on Safari. What started as a quick internal hack slowly grew into its own project.

Open sourcing SnapDOM changed everything. Suddenly the audience wasn’t just me, but anyone with their own browser, CSS, and HTML quirks. That’s when I realized how immense the attack surface really is. Every combination of engines, DOM structures, styles, and external resources can break in unexpected ways. At times it felt overwhelming, but I’ve learned to treat each bug as a reproducible pattern: isolate it, fix it, and carry the lesson forward.

From the beginning, two principles guided the work: fidelity and speed. Capturing “every detail” would not matter if the process was unbearably slow, so I spent a lot of time on caching strategies, incremental updates, and efficient DOM cloning. Benchmarking against existing libraries became part of the routine. Performance was never an afterthought — it was always as important as accuracy.

On the technical side, SnapDOM pushed me into DOM cloning strategies, caching, font embedding heuristics, browser-specific hacks, and the constant tradeoffs between fidelity and performance. On the personal side, it has been just as educational: patience, humility, and learning to really listen to users. Some of the best moments have been receiving thoughtful PRs and clear reproduction cases — collaboration that genuinely made the project better. The respect shown in these contributions has been one of the most motivating parts of the journey.

Another milestone: only a few months in, SnapDOM already has its first two GitHub Sponsors and has passed 6k stars on GitHub. Knowing that people found enough value to sponsor the project, and that thousands more cared enough to star it, has been incredibly rewarding.

I’ve also been designing a plugin system, so the core stays lean while features like cropping, filters, or alternative exporters can live as extensions. It’s not merged into the mainline yet, but it’s a direction I’m excited about — and another layer of complexity to manage as the project evolves.

It’s still early days, but reflecting on the journey from April to now has been meaningful. What started as a small internal helper for a zoomable UI turned into an open source project with real users, contributions, sponsors, and community respect. Balancing the technical struggles with the human side of open source has made this one of the most rewarding experiences I’ve had so far.

Thank you, Martin

Repo: https://github.com/zumerlab/snapdom