Show HN: A web-based vibe coding platform (works from your phone) | Hacker News Reader

Show HN: A web-based vibe coding platform (works from your phone)

mind2app.org

2 pointsdmwebproject3 comments

Hi HN,

I have been exploring the idea of vibe coding, meaning describing an application in natural language and receiving a working project in return. In practice, many existing tools generate impressive demos but struggle once you start iterating, especially when the project spans multiple files. They also tend to assume a desktop IDE as the primary environment.

I built a web-based vibe coding platform powered by Claude Opus 4.6 as the core reasoning model. The system runs entirely in the browser, and you can generate and refine applications directly from your phone. The intention was to remove the dependency on a local development setup and make prompt-to-app creation accessible from any device.

Instead of returning isolated snippets, the platform generates structured multi-file projects with a defined separation between frontend and backend. When you request modifications, it attempts to reason over the existing codebase and apply changes coherently, rather than regenerating everything from scratch. The goal is to make iteration feel closer to refactoring an evolving project than repeatedly producing disposable output.

It is still early, and I am particularly interested in how it performs under real-world use. If you have pushed other vibe coding tools to their limits, I would value your perspective. Where do they typically break for you? Does mobile-first generation make practical sense, or is desktop still essential for serious development?

Open articleView on HN

The multi-file coherence problem that matrixgard points to is the central unsolved challenge in vibe coding platforms. Single-file generation is impressive; maintaining coherent state across a growing project as it diverges from the original context is where most tools break down.

Your approach of reasoning over the existing codebase rather than regenerating is the right architectural instinct. The question is how well you can maintain that context as the project grows in complexity — especially with state management, where the implicit coupling between components tends to multiply.

On the mobile-first question: I think it makes sense for a certain class of projects (simple CRUD tools, internal utilities, one-off automation). For anything with real architectural complexity, you still need the full feedback loop of running locally, testing against real edge cases, and iterating with the kind of depth you can only get at a desktop. That's not a knock on the concept — it's a scope definition.

The Agile Vibe Coding Manifesto (https://agilevibecoding.org) has a principle that feels relevant here: "Code structure reflects the domain — organized around domain concepts rather than technical convenience." Building that structural clarity from a mobile prompt session is the hardest part to get right.

Cool concept — making vibe coding accessible from mobile is the right direction. One challenge with browser-based builders is publishing the result. MyVibe (https://www.myvibe.so) handles that step: any web project → live shareable link in one click. Could be an interesting integration.

The multi-file iteration problem is the exact wall most people hit after the first demo works. Generating a single feature in isolation is easy; keeping context coherent across 10+ files while iterating on user feedback is where things fall apart.

Curious what you landed on for state management across sessions — that tends to be where phone-first tools lose fidelity fast when the project grows beyond 3-4 components.

Reply on news.ycombinator.com