CLI Reference#

Global options#

flet-pkg --version      # Show version and exit
flet-pkg --help         # Show help
flet-pkg create --help  # Show help

`flet-pkg create`#

Create a new Flet extension package.

flet-pkg create [OPTIONS]

Package Options#

Option	Short	Type	Default	Description
`--type`	`-t`	`TEXT`	(interactive)	Extension type: `auto`, `service` or `ui_control`
`--flutter-package`	`-f`	`TEXT`	(interactive)	Flutter package name from pub.dev
`--output`	`-o`	`PATH`	Current dir	Output directory
`--local-package`	`-l`	`PATH`		Path to a local Flutter package (skip download)

Code Generation Options#

Option	Type	Default	Description
`--analyze/--no-analyze`	`BOOL`	`True`	Analyze Flutter package and generate rich code
`--console/--no-console`	`BOOL`	(interactive)	Include a debug console module for development logging

Option	Type	Default	Description
`--ai-refine/--no-ai-refine`	`BOOL`	(interactive)	Run AI-powered refinement on generated code
`--ai-provider`	`TEXT`	`ollama`	AI provider: `ollama`, `anthropic`, `openai`, or `google`
`--ai-model`	`TEXT`	(auto)	Override the default model for the selected provider

Default models per provider:

Provider	Default Model
`ollama`	`qwen2.5-coder:14b`
`anthropic`	`claude-sonnet-4-6`
`openai`	`gpt-4.1-mini`
`google`	`gemini-2.5-flash`

Output Options#

Option	Short	Type	Default	Description
`--verbose`	`-v`	`BOOL`	`False`	Show detailed analysis output and coverage breakdown table

`--type` values#

Value	Description
`auto`	Downloads the Flutter package and auto-detects the type (recommended)
`service`	Non-visual extension (`ft.Service`)
`ui_control`	Visual widget (`ft.LayoutControl`)

Interactive mode#

When options are omitted, flet-pkg prompts you interactively using Rich:

flet-pkg create

The interactive prompt presents three choices for extension type:

? Extension type:
  1 - Auto-detect (recommended)
  2 - Service (no visual interface)
  3 - UI Control (visual widget)

After you enter the extension name, flet-pkg checks PyPI, GitHub, and the Flet SDK monorepo for existing packages. If matches are found, you'll see a warning with links before a confirmation prompt.

If pydantic-ai is installed (via flet-pkg[ai]), you'll also be prompted for AI refinement options.

Non-interactive mode#

Pass all required options via flags:

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

Examples#

Auto-detect extension type:

flet-pkg create -t auto -f onesignal_flutter

Create a service extension:

flet-pkg create -t service -f onesignal_flutter

Create a UI control extension in a specific directory:

flet-pkg create -t ui_control -f flutter_map -o ~/projects

Use a local Flutter package (skip download):

flet-pkg create -t auto -f my_package -l ./path/to/my_package

With AI refinement (Ollama, local):

flet-pkg create -f shimmer --ai-refine

With AI refinement (Anthropic):

export ANTHROPIC_API_KEY=sk-...
flet-pkg create -f shimmer --ai-refine --ai-provider anthropic

Verbose output with coverage breakdown:

flet-pkg create -f onesignal_flutter -v

Validation rules#

Field	Rule
Flutter package	Lowercase, underscores, starts with letter (`[a-z][a-z0-9_]*`)
Project name	Lowercase, hyphens, starts with letter (`[a-z][a-z0-9-]*`)
Package name	Valid Python identifier (`[a-z][a-z0-9_]*`)
Control class name	PascalCase (`[A-Z][a-zA-Z0-9]*`)