Most AI app builders jump straight into building. That feels fast until your app gets messy, inconsistent, and nothing like what you imagined. Today, we’re introducing Spec Coding in Capacity. Instead of building immediately, Capacity now lets you define your app first with the help of an AI co-founder that asks the right questions before any code is generated. More structure upfront. Far less refactoring later. Much better results.
Hey Product Hunt 👋
Samuel here, maker of Capacity.
We love vibe coding it’s fun and incredibly fast.
But after building a lot of real apps with AI, we kept running into the same problem:
AI builds confidently, even when the idea isn’t clear.
Humans ask questions.
AI doesn’t, it guesses.
So we built Spec Coding.
Before any code is written, an AI co-founder helps you define:
-what you’re building
-who it’s for
-what matters
-and what doesn’t
It slows down the first minute…
and saves hours later.
Would love your feedback especially if you’ve ever rebuilt the same AI app three times 😄
Happy to answer any questions!
A really good idea! Apps like this need more business analysis, which includes quite a few important steps. Only after completing those steps should the code be generated. That’s when it becomes truly useful.
My previous AI projects always turned into a mess after a few prompts. This "plan first" approach is a lifesaver for someone who isn't a pro developer.
$69/month for even one custom domain? Industry standard is $25, and Dyad.sh allows one even on their free plan.
Direct Github integration cxoming soon...?
Also, as for the problem with AI "not asking questions," this is why we RAG it: we tell our AI tools, "when something is unclear, always ask questions before you answer or build."
Vibe Mode / Spec Mode: I read your explanations, and now have absolutely no idea which to choose. You need a help page with example(s).
Finally, I did get in there, found the demo does not let me enter a prompt; doesn't work.
Oh, bnut the Mini Pack, for when we get stuck and need just a few more credits: brilliant!
I like the co-founder concept—does it ask clarifying questions at each step, or does it front-load all the planning before generating anything?
The spec-first approach resonates—I've seen plenty of vibe-coded apps where early assumptions snowball into major rewrites. I'm curious how the system handles evolving requirements mid-project. Does the AI co-founder flag when new requests conflict with the original spec, or does it silently adjust?
Been following you from far on X, liked the Starter Story video too
Cool launch and very nice product you guys cooked here, congrats for top 3 :)
The "planning first" approach addresses the biggest pain point in vibe coding – fast initial iterations that turn into technical debt. Most AI builders optimize for speed-to-first-render, not maintainability.
This spec-first flow reminds me of design systems thinking: define constraints upfront, iterate freely within them. Does the AI co-founder maintain spec consistency across revisions? That continuity would be critical for teams building beyond MVPs.
Really interesting approach with Spec Coding! As a full-stack dev building SaaS apps (DeadlineKeeper for college applications), I've been using BMAD Method (Breakthrough Method for Agile AI Driven Development) with Claude Code. The idea of defining structure upfront with an AI co-founder resonates with me - BMAD has 21 specialized AI agents that help architect before coding. Will definitely check out Capacity to see how your spec coding compares! 🚀
Having used other vibe coding tools, excited to try another method to get started! Is this already connected on the backend to a DB, and does it support something google login, or payment systems?
Thanks for sharing this @samuel_rondot ! Any free coupons for us PH users? 🤓
This looks great and really helps put the focus on the most important stage of the process. Do you have features to allow teams to work together on this? Good luck with the PH launch.
Nice, 2 questions: 1. How much it differs from planners such as the one that Cursor and other tools have (when they set a long todo list) + Planning mode 2. How consistent the agent stays on the created initial plan? and how flexible it can be to adjust the plan and micro-steps based on progress it makes?
does the co-founder assistant allow me to export the technical documentation separately if I want to keep a record of the app structure?
I really appreciate the focus on reducing refactoring because I usually spend more time fixing AI-generated messes than actually building the features I want. 🛠️
Slowing down at the start to save time later feels obvious now, but I never thought of it like this. Do you guide users on what “matters” vs “doesn’t matter”?
