Peter Steinberger, a well-known name in the developer community, showed how a single hour of vibe coding can produce a viral product and turn a career upside down. His OpenClaw story is the quintessence of what we call "vibe coding turned up to eleven."
Who he is
Peter Steinberger is a developer who, in November 2023, sitting in Marrakesh with friends, wondered: "What if I just wired WhatsApp up to the Cloud Code CLI?" That idea became the starting point for a prototype that later turned into OpenClaw (originally known as Clawd's, MoldBot, WA-Relay). His approach had nothing to do with traditional planning and months of development. He just went and built it.
How it was done
It started with an hour of coding. Peter connected WhatsApp to the Cloud Code CLI, and the prototype worked: the agent ran commands on his Mac and answered in chat. It wasn't just fast, it was intuitive and effective. He didn't plan every feature up front — he added them as the need or the inspiration showed up. Image support first, then voice messages. When he simply dropped in an audio file with no extension, the agent figured out on its own that it was opus, converted it with ffmpeg, sent it to OpenAI Whisper and returned the text. That shows a deep feel for systems and the ability to use what they offer without extra effort.
The hardest part, according to the source, was that Peter made the agent "self-aware." Not in a philosophical sense but a practical one: the agent knew its own source code, its documentation and the model running it. That let it be remarkably flexible and autonomous. This way of building — creating a product in flow, with minimal planning and maximum output — is a vivid example of vibe coding. It didn't just save time, it let the product evolve organically, answering real needs and real opportunities.
The numbers
- Time to build the prototype: 1 hour.
- GitHub stars: 180,000 in three months.
- Recognition: the fastest-growing repository in GitHub history.
- Outcome: Peter Steinberger was hired by OpenAI.
What to take away
- Move fast: don't be afraid to start with a minimal prototype. One hour of coding can tell you more than months of planning. If you have an idea, try to ship it as fast as possible to test whether it holds up.
- Use what already exists: Peter didn't build everything from scratch, he "wired WhatsApp up to the Cloud Code CLI." Integrating existing platforms and services speeds development up dramatically and lets you focus on your product's unique value.
- Let the product evolve: instead of a rigid plan, give the product room to grow organically, adding features as needs or new ideas appear. Flexibility is the key to building something genuinely innovative.
- Aim for "self-awareness": think about how your product can "know" itself — its code, its documentation. That doesn't just simplify maintenance, it opens up new options for automation and self-optimization.