Twitch Highlights

Tzelon Machluf and Ron were two 30-year-old developers from Israel who'd been friends since childhood. Like many technical co-founders, they had started and abandoned numerous projects over the years, always waiting for "the right time" to go all-in. As enthusiastic Twitch viewers juggling families and full-time developer jobs, they faced a real frustration: streams averaged ~8 hours, making it nearly impossible to catch everything. They took inspiration from sports—specifically how NBA publishes highlight reels minutes after each game—and wondered why Twitch streamers didn't have the same option.

Following *The Lean Product Playbook*, they did early validation: created personas, refined features, built a landing page, and sent cold emails to professional streamers. They got a few positive responses and many silences, but enough signal to justify the leap. They quit their jobs and went all-in. For three months, they wrote code nonstop, learning Computer Vision and OCR from scratch. They used Node.js and opencv4nodejs (the library creator, Vincent Mühler, mentored them). Their breakthrough was combining multiple signals: detecting text overlays in games (like "Victory Royale" in Fortnite), measuring anomalies in Twitch chat message frequency, and other "secret indicators" to pinpoint exciting moments. They reduced editing time from ~12 hours to ~4 hours per stream—a real win.

Three months in, they hit a wall: the streamers who had responded to their initial cold emails stopped answering. They couldn't find beta testers in the "streamers world" or through other social networks. This was their first obvious failure, and it signaled a deeper problem: they had no audience, no community, no leverage.

What worked: the technical execution. They built something genuinely useful that solved a real problem. What failed catastrophically was everything else. They didn't understand the importance of marketing or building an audience *before* going all-in. Cold emails and Reddit posts got zero traction for an unknown brand. When they couldn't find beta testers, motivation evaporated. Instead of confronting this core problem—the lack of audience—they added more features ("classic mistake"). After 8 months and ~$2,500/month in living expenses plus ~$500/month on AWS, they ran out of money with zero customers and zero momentum. The product was good. The market validation was nonexistent.

Tzelon went on to research "How to launch a product as a Maker" and built CreateCamp.co to help other makers avoid the same trap. Both founders extracted a crucial lesson: building an audience and reputation in your target community takes months or years *before* you build the product. Ron emphasized the need for more networking and community research upfront. Tzelon's key insight: treat it like a marathon, not a sprint. Build your platform first, establish trust, *then* launch the solution.