Skip to content

RetroDash

Plataforma de Retrospectiva

Retrospectivas melhores, a cada sprint.

Um quadro de retrospectiva privado e em tempo real para equipes Scrum e Kanban. Reflita com honestidade, vote juntos e defina ações que realmente acontecem.

retrodash.app/room/sprint-42Live · 4

What went well

To improve

Action items

Type
Product
Year
2026
Status
In production
Role
Solo project: concept, design, engineering and deployment
Stack
Next.js, React, TypeScript, Tailwind CSS, Firebase, Vitest, GitHub Actions
Visit the site (opens in a new tab)

RetroDash is a real-time tool for agile retrospectives. The team opens a room, writes cards, votes and leaves the meeting with action items. It’s free and doesn’t ask for a credit card.

I built all of it on my own: concept, visual identity, interface, business rules, data model, architecture, tests, CI/CD and deployment.

What it does

  • Public rooms, or private rooms with a password.
  • A real-time board: cards, votes and status show up for everyone in the room without reloading.
  • Anonymous mode, which hides who wrote each card, on the leaderboard too.
  • Action items with three states (Pending, Done and Keep) that can carry over to the next room.
  • A podium with gold, silver and bronze: 1 point per card and 0.5 per action item.
  • “Improve with AI”, which rewrites a card or an action item of up to 500 characters.
  • Interface in Portuguese and English.

Engineering decisions

  • Real time comes from Firestore listeners. Every card, vote and status change reaches everyone in the room as it happens.
  • Organizations, rooms and invites are multi-tenant, with roles and server-side validation.
  • Private room passwords are stored as SHA-256 hashes.
  • “Improve with AI” calls Gemini Flash through a server-side API, so the key never reaches the browser. Cards and action items have separate modes.
  • The design system is my own, built with Tailwind and CSS variables, with light, dark and system themes.
  • Tests with Vitest, a GitHub Actions pipeline and releases published with semantic release.

Message me on LinkedIn (opens in a new tab)