RetroGemini

RetroGemini: open-source retrospectives and team health checks for internal use I'm the maintainer of RetroGemini, a free and open-source application for Scrum retrospectives and team health checks. I developed it in my work at the State of Geneva because we needed to keep potentially sensitive team feedback on our own infrastructure. Around 30 teams in my workplace now use it, with encouraging feedback. It is released under The Unlicense. Teams can collect ideas in real time, group cards, vote, track actions across sessions and run team health checks. The interface is available in English and French. You can start with Docker and SQLite, or use PostgreSQL and the Kubernetes/OpenShift deployment guides. The core app can run on an isolated network; optional AI features require explicit configuration. Source and installation instructions: republique-et-canton-de-geneve/RetroGemini It is also listed in the Swiss OSS catalogue: opensource.admin.ch AI transparency: Claude and Codex assisted substantially with development. AI also helped draft this post. I'd welcome feedback on installation and on what would make the application useful to other teams or public administrations.

RetroGemini: self-hosted Scrum retrospectives used by around 30 teams at the State of Geneva I'm the developer of RetroGemini. I work at the State of Geneva, where we needed to keep potentially sensitive retrospective feedback on our own infrastructure rather than put it in external cloud tools. I built an alternative to cloud tools with AI assistance. RetroGemini covers collaborative retrospective boards, grouping and voting, action follow-up, and team health checks. The interface is available in English and French. For self-hosting: - Docker with persistent SQLite storage for a simple setup. - PostgreSQL and Kubernetes/OpenShift deployment options. - Static assets hosted locally, so the core app can run on an isolated network. - Released under The Unlicense. Repo and installation instructions: republique-et-canton-de-geneve/RetroGemini For a local evaluation: docker run -d --name retrogemini -p 8080:8080 -v retro-data:/data jpfroud/retrogemini:latest Then open localhost. The README covers the stable session secret and HTTPS configuration for a production deployment. AI disclosure: Gemini, Claude and Codex assisted substantially with development. The core application does not require an AI service; optional OpenAI-compatible features are off unless configured, and their data flow depends on the endpoint you choose. If you run retrospectives on internal infrastructure, I'd be interested in what makes a tool practical for your team, and in feedback from trying the installation.