Success Stories
The framework is real, battle-tested at small scale. Here is what one team built when they took it seriously.
Below is one team's reference implementation, drawn from their own retrospective. It is what DoCoDeGo looks like when one team takes it seriously end-to-end. We are in Alpha; external case studies are forthcoming as partner teams ship.
Shipped a project using DoCoDeGo? Tell us — /collaborate. We will write it up with you, credit your team, and link to your repo.
docodego-starter
A full-stack reference implementation built from spec-first principles: 7 application surfaces, 7 shared packages, one design system, one auth layer. Every feature traceable from spec to production behaviour.
- Web Next.js storefront
- Mobile React Native client
- Desktop Electron shell
- API Node backend
- Browser ext WXT extension
- VS Code Editor extension
- CLI TypeScript command-line
A single shared design-system package shortened cross-app iteration to hours, not days. A change to a primitive propagates to every surface in one composition cycle.
All four pillars, observable in practice.
No invented scorecard. What follows is what the team actually did at each pillar, drawn from the project's own retrospective.
Every feature shipped with a structured spec; specs live alongside the code they generated.
Multi-agent workflow with explicit architectural sign-off before each cycle.
Validation-gated CI; deployments triggered by passing acceptance criteria, not the calendar.
Weekly Intent Reviews, drift tracking, kill-switch authority held by a named human.
Four things the data told us.
- ▸ Specs scored ≥ 80 produce AI output that ships with no architectural rework.
- ▸ A single shared design-system package shortened cross-app spec iteration to hours.
- ▸ Multi-agent composition is faster than single-agent for cross-cutting changes — the bottleneck moves to architectural review, not generation.
- ▸ Intent Reviews caught drift in week 6 that would have surfaced as a production incident in week 12.
More success stories coming.
These patterns came from real teams. Discuss your own implementation in #success-stories on Discord — case studies and reference starters land there first.
The framework is real.
The community is forming now.
DoCoDeGo is in Alpha. The framework is documented, the practices are battle-tested at small scale, and the next release is being shaped in public.
If it produces anything, it should produce engineers and teams who think more clearly about what they are building and why.
Discord is where specs are debated, the framework gets sharper, and decisions land in writing. The conversation is the artefact.