Plugin development tutorials

This is the hands-on path for building a MINT plugin. Start with a small analysis plugin, add a browser frontend, then build a design plugin with tables and role checks.

For the full Plugin Development chapter map, start at Plugin Development Guide.

Before you start

You'll need:

Python	3.12 or newer
uv	Used by generated plugin projects
Bun	latest (for frontend tutorials)
mint CLI	Installed from mint-sdk; see CLI overview
A running platform	Optional for the first pass; mint dev --platform can start one for integrated smoke tests

You do not need to read the whole SDK reference first. When a tutorial introduces a concept, it links to the deeper page.

Recommended path

Step	Tutorial	Project	What you will have afterward
1	First analysis plugin	hello-mint	Backend-only ANALYSIS plugin with routes, tests, PlatformContext reads, and a .mint bundle
2	Adding a frontend	hello-mint	Vue 3 frontend using PluginWorkspaceView, generated typed client, and @morscherlab/mint-sdk
3	Design plugin with tables	panel-designer	EXPERIMENT_DESIGN plugin with SQLModel tables, CRUD routes, local SQLite, and installed-mode migrations
4	Plugin roles	panel-designer	Plugin-specific viewer / editor / admin roles enforced by backend routes and reflected in the UI

The first two tutorials build one analysis plugin. The last two build a separate design plugin. That split mirrors real MINT deployments: design plugins usually define experiment structure, while analysis plugins read experiments and attach results.

What to learn before each step

Before	Read this if the term is new
Tutorial 1	Plugin types, PlatformContext
Tutorial 2	Frontend overview, Frontend SDK reference
Tutorial 3	Data model, Migrations
Tutorial 4	Route permissions, Permissions

How tutorials are structured

• Command blocks are runnable as written. Code blocks are either full replacements or explicitly marked as partial snippets.
• File paths are absolute relative to the plugin project root unless noted.
• When directory matters, the surrounding text tells you where to run the command.
• Output expected from each command is shown with → markers.
• After each section there's a Checkpoint — a one-liner you can run to verify your project is in the expected state before moving on.

After the tutorials

When your plugin works locally:

• Run Packaging to produce a .mint bundle.
• Use Publishing when you are ready for PyPI or a marketplace registry.
• Add CI from CI patterns.
• Use Recipes for specific tasks such as reading experiments, writing results, route permissions, testing, and R integration.
• Use API Reference only when you need exact signatures.