flet-pkg

CLI tool to scaffold Flet extension packages.

Generates complete, ready-to-develop Flet extensions with Python + Flutter/Dart code, following Flet 0.80.x+ patterns.

Important: this tool does not perform miracles. 🙏😂

Every Flutter package on pub.dev has its own API surface, configuration requirements, and platform-specific behaviors. Some packages expose simple method calls; others require complex initialization flows, native platform setup (Android manifests, iOS plists, Gradle/CocoaPods configuration), or callback-based architectures that don't map cleanly to Python.

It would be unrealistic to write a code generation algorithm that generalizes perfectly for every possible Flutter package. flet-pkg aims to accelerate the process by delivering significantly more than a blank skeleton — it downloads the real Dart source, parses the public API, and auto-generates Python controls, type mappings, enum definitions, event handlers, and Dart bridge code. For service packages, this typically covers ~95% of the API surface.

However, the developer is still responsible for:

Understanding how the original Flutter/Dart package works
Reviewing the generated code for correctness
Handling package-specific configurations (native platform setup, initialization flows, etc.)
Adjusting or complementing the generated code for edge cases the pipeline could not cover

Think of flet-pkg as a smart starting point, not a finished product.

Features#

Interactive & non-interactive modes — guided prompts or full CLI flags
Auto-detect extension type — downloads the Flutter package and detects whether it's a Service or UI Control automatically
Three extension types — Auto-detect (recommended), Service (non-visual), and UI Control (visual widget)
Name conflict detection — checks PyPI, GitHub, and the Flet SDK monorepo before creating, warns if the name already exists
Complete project scaffolding — Python package, Flutter package, tests, docs, examples
Smart name derivation — automatically derives project, package, and class names from the Flutter package name
Flet 0.80.x+ patterns — uses @ft.control, ft.Service, ft.LayoutControl, invoke_method, and EventHandler
AI refinement — optional LLM-powered code improvement using the Architect/Editor pattern
MCP Server — expose scaffolding and analysis tools to AI agents via the Model Context Protocol

Quick install#

pip install flet-pkg

Or with uv:

uv tool install flet-pkg

Quick start#

flet-pkg create

This walks you through creating a new Flet extension with interactive prompts.

For non-interactive usage (auto-detect type):

flet-pkg create --type auto --flutter-package onesignal_flutter --output .

Or specify the type explicitly:

flet-pkg create --type service --flutter-package onesignal_flutter --output .

uv add flet-pkg[ai]
ollama pull qwen2.5-coder:14b
flet-pkg create -f shimmer --ai-refine

Generated project structure#

flet-onesignal/
├── pyproject.toml
├── README.md
├── CHANGELOG.md
├── LICENSE
├── mkdocs.yml
├── docs/
├── tests/
├── examples/
│   └── flet_onesignal_example/
└── src/
    ├── flet_onesignal/        # Python package
    │   ├── __init__.py
    │   ├── onesignal.py       # @ft.control + ft.Service
    │   └── types.py
    └── flutter/
        └── flet_onesignal/    # Flutter package
            ├── pubspec.yaml
            └── lib/
                └── src/
                    ├── extension.dart
                    └── onesignal_service.dart