Between what you mean
and what gets built,
something gets lost.
Not because AI isn't capable.
Because intent was never structured before building started.
You describe something to Claude Code. It builds. The next session, it builds from where it thinks it left off — which isn't quite where you left off. By the time the drift is visible, rewinding is harder than starting over.
The gap is not a prompt problem. It is a structural one. Your intent was never written down in a form that persists. It lived in a chat window. The chat window closed.
Most tools address this after the fact — better prompts, auto-memory, clearer mid-build instructions. These help. They don't fix the root.
Unstructured intent, suspended. Ada compiles it into something that governs.
Claude Code's auto-memory records what it observed as it worked.
Ada records what you intend before it starts.
One is retrospective. One is prospective. They are not the same thing.
Ada is the step before.
You describe what you want to build — at whatever level you think in. Ada asks you the questions that surface what you actually mean. Not framework questions. Not library questions. Questions about what the thing should do and be.
You answer. Ada compiles. The files land on disk.
Less distance between what you described and what gets built.
Not zero. Less.
What Ada produces
CLAUDE.md Claude Code reads this on every message turn — not just at session start. What the project is. What it is not. What constraints apply. Injected into every context window. agents/ Specialized agents scoped to bounded contexts. Each knows its domain and its constraints. hooks/ Pre-tool guards that run before Claude Code takes any action. The constraints from your intent, enforced at the boundary where decisions are made. Plain files. You read them before handing to Claude Code. If they match what you meant: proceed. If not: run Ada again with the correction.
git clone https://github.com/alexrozex/ada Early access — npm package coming soon.
Built by one person. ~400 iterations since late 2024.
Read about Ada →