Notes from the shipping lane.
Practical writing on AI app generation, codebase-aware tooling, and getting from prompt to production.
Shipd: an AI app builder for real applications
What Shipd is and how it works: an AI development team, codebase-aware generation, feature-by-feature builds, live preview, and a workflow that continues after version one.
How the AI team actually builds a full-stack app
Inside Shipd's AI team pipeline: how context passes between five specialists, which Claude model runs the coding roles, and how the frontend gets built against a real API.
AI that understands your codebase: how Shipd works with real projects
How Shipd's codebase-aware workflow actually works: analyse the repo, find where a change belongs, reuse your components, and keep building the same app.
Meet your AI team: PM, Director, Designer & Engineers
How does an AI development team work? Meet the PM, director, designer, and engineers who turn ideas into real products and ship them.
AI for coding: from code generation to codebase-aware development
AI can generate code from a prompt. The harder question is whether it can understand your existing application first. Why codebase-aware AI is the next phase of AI for coding.
The AI team now ships the backend too. And builds on our top model
Shipd's team gained a backend engineer and moved its coding roles onto Claude Opus 5. A PM still interviews you, a designer still sets the direction. But what lands at the end is now a full-stack app, built feature by feature.
Why we stopped generating whole apps in one shot
One prompt, one giant generation, and you debug whatever comes out. Shipd now plans the features, builds them one at a time, and checks each one in a live preview before moving on. So what reaches you already works.
Why we built an AI team instead of another AI chatbot
Every AI app builder is one model in a chat box. Shipd runs a visible team of specialists, PM, director, designer, and two engineers, who plan, design, build, and review your app while you watch.
What "full-stack" should mean in an AI app builder
Most AI builders say full-stack and mean a frontend wired to someone else's database. Here's the difference between hooking up a BaaS and generating a real backend that runs and deploys.
Codebase-aware, part two: staying in sync with a repo that keeps moving
Understanding a repo once is the easy part. Here's what we shipped to keep that understanding fresh as your code changes, your project spans more than one repo, and your whole AI team needs to read the same architecture.
How to move your Lovable, Bolt, or v0 project into Shipd
Greenfield AI builders start you over every time. Here's how Shipd imports the project you already have, your real code and assets, so you can keep building it instead of rebuilding it.
What you can actually build on Shipd's free tier
Shipd's free plan gives you 100k tokens a month, no card required. Here's an honest look at what that covers and where the ceiling shows up.
What "codebase-aware" AI generation actually means
Importing a repo isn't the same as understanding it. Here's the difference between connecting a codebase and generating code that actually fits it.
How to export a working web app to editable Figma layers
Figma-to-code is everywhere. The reverse direction, turning a live app into editable Figma layers, is rarer, and more useful than it sounds. Here's how it works in Shipd.