minddy runs the interface
The Next.js application serves the tracker and its API.
Start by installing the desktop app. It will open the same public minddy core on the local folder or server you control.
Self-hosted minddy uses the public release. You choose the infrastructure and take responsibility for its operation.
01
The desktop app opens the minddy interface from your chosen data source. Behind it, the self-hosted Next.js service and Supabase provide the tracker, accounts, attachments, and live updates.
The Next.js application serves the tracker and its API.
Auth, Postgres, Storage, and Realtime work as one stack.
Projects, accounts, and attachments stay on the infrastructure you choose.
PostgreSQL alone is not supported. minddy also needs Supabase Auth, Storage, and Realtime.
02
Every route begins with the desktop app, then uses the same open-source release. Pick by audience: just you, or a whole team.
A private local instance for one developer, managed from the desktop app with no domain or server administration.
Run a shared instance on a VPS or a machine at home. Start on your private network without a domain, or make it public with HTTPS.
03
Your data doesn't have to stay behind. Export it from Cloud as one JSON file, install your own instance with the wizard, then import the file there — projects, issues, comments, and attachments come along.
1. Export from Cloud
Settings → Data → “Download my data”. You get a minddy-export-….json file with no passwords or keys inside.
2. Import on your instance
Settings → Data → “Import a transfer file”. The import is additive: it keeps original identifiers when they are free and reports whatever it had to remap.
Credentials never travel in the file: reconnect GitHub, GitLab, and AI providers inside the app after importing.
The public core includes the tracker, administration, exports, MCP, and documented agent paths. AI, email, Git, analytics, push, and external storage are optional: select and configure only the providers you want to use.
Cloud billing and product analytics are not prerequisites. With only required settings, the instance does not call Minddy Cloud, Stripe, PostHog, or an AI provider.
Keep the runbook from your deployed release. These actions are part of running a shared instance.