How an image-heavy marketplace scaled to 4M+ images and 50TB monthly traffic
- industryEcommerce, Marketplace
- founded2021
- headquartersZurich, Switzerland
- marko.ch
Key results
- 6 weeks of engineering time saved before launch
- Media infrastructure supports 4M+ images, 10 TB stored, and 50 TB monthly traffic
- Faster, more responsive product experience through automatic WebP and AVIF delivery
It just works out of the box without having to think about it. That’s the most important part for us.
marko is Switzerland’s community-driven secondhand marketplace, built for Gen Z and millennial consumers who care about sustainability, affordability, and style. Because every listing depends on photos, image uploads and delivery sit at the center of both seller activation and buyer conversion.
As a lean startup preparing for launch, marko initially explored building its own image pipeline. That effort quickly proved too complex for the timeline. Uploadcare gave the team a reliable media layer for uploads, transformations, delivery, and cleanup, helping them launch on time, keep the product fast, and support AI-assisted listing flows from day one.
marko is a community-driven secondhand marketplace where users buy and sell items through image-based listings. The product depends on user-generated images. Sellers need to upload photos quickly, buyers expect crisp listings that load instantly, and those same images also power AI-assisted listing flows that prefill titles, categories, brands, colors, and pricing suggestions. That makes media infrastructure a core part of the product rather than a background service. With a fixed launch window approaching and major dependencies still in progress, there was a clear need for reliable image handling that wouldn’t slow development.
Initially the team tried building its own CDN image pipeline, but the scope expanded quickly. Uploading files was only the beginning. They also needed multiple output sizes, retina-friendly rendering, on-the-fly transformations, deletion logic for abandoned listings, and reliable delivery at scale. For a small engineering team, continuing down that path would have delayed launch by roughly another month and pulled resources away from product-critical work.
We realized it was more than we expected. Uploading was one thing, but handling transformations, delivery at scale, and everything around it just wasn’t in our best interest to build ourselves.
marko adopted Uploadcare early as the foundation of its media pipeline. Rather than combining separate tools for storage, CDN delivery, transformations, and cleanup, the team moved those components into one infrastructure layer that could scale with the marketplace.
With marketing plans, launch dependencies, and a tight engineering schedule already in motion, marko needed a solution it could implement quickly instead of extending the roadmap for a non-core system. By choosing Uploadcare before launch, they reduced the scope the team had to carry internally and made it possible to stay on track for go-live.
Instead of having a complex set of media requirements, Uploadcare gave marko a stable system that could be reused across the product.
- End-to-end media pipeline: Uploadcare provides uploads, storage, transformations, and CDN delivery in one system, preventing the team from diverting resources to build a custom pipeline.
- Dynamic image transformations: Images can be resized and adjusted on demand through URL parameters, allowing the team to serve the right dimensions and formats across app and web interfaces.
- Automatic format optimization: Images are delivered in optimized formats such as WebP and AVIF, helping reduce file sizes while maintaining visual quality.
- Image lifecycle management: Uploadcare enables automatic cleanup of images from abandoned listing flows, preventing unnecessary storage and removing the need for custom deletion logic.
We’re not in the business of image transformation. We’re in the business of secondhand sales.
With media infrastructure handled by Uploadcare, marko could focus on launching and scaling the marketplace instead of maintaining image systems.
- Faster path to launch: Avoiding a custom-built image pipeline saved roughly six weeks of engineering time, keeping the team on schedule for their marketplace launch.
- Engineering focus on core features: Instead of building and maintaining media infrastructure, the team concentrated on critical functionality such as shipping integrations, admin tooling, and the marketplace purchase flow.
- Marketplace scale supported: marko supports 300,000 monthly users, over 500,000 listings, and 4+ million images, with around 50 TB of image traffic per month—all powered by a scalable media pipeline.
- Foundation for AI-powered listing workflows: Reliable and instantly accessible media enables marko’s AI features to analyze uploads and prefill listing details, helping sellers create listings faster.
For a startup like marko, where every engineering hour counts, having a reliable media layer that we don’t have to think about lets us focus on building the features that differentiate us.
Handling files is hard.
We make it easy.
Read similar stories
