AI NewsDev toolsAnnouncement

Flet 1.0 builds Python apps for iOS, Android, desktop and web

Flet 1.0 is out after more than four years of work, letting Python developers build one codebase that runs as an app on iOS, Android, Windows, macOS, Linux and the web, with the framework at 16,858 stars and more than 9 million PyPI downloads.

AI News

Editorial2 min read

LinkedInX
Flet repository page on GitHub

Image: GitHub

Why it mattersA Python developer with no frontend background can now ship a real mobile or desktop app from one codebase, with test tooling that runs the packaged app on a phone and a stated compatibility policy for future upgrades.

Appveyor Systems has released Flet 1.0, the first production version of a framework that lets Python developers build one codebase that runs as an app on six platforms: iOS, Android, Windows, macOS, Linux and the web. The 1.0 announcement was posted on 17 September 2026 and reached 79 points on Hacker News overnight. Flet says the release means the framework is ready for building production apps.

The project has been public for more than four years. The GitHub repository is at 16,858 stars, with 100 contributors, and Flet has been downloaded from PyPI more than 9 million times. The team has merged over 1,000 pull requests, closed 2,800 issues, and its Discord community has grown past 6,000 members. The framework is Apache-2.0 licensed and Flet is written in Dart on the Flutter side and Python on the SDK side.

What is new since the alpha

Flet says the run-up to 1.0 focused on testing. Framework unit tests run across Python 3.10 to 3.14. Control and example integration tests compare screenshots to catch visual regressions. Mobile package integration tests exercise native libraries on Android and the iOS simulator. flet build runs an end-to-end matrix of Python 3.12, 3.13 and 3.14 across Windows, macOS, Linux, iOS, Android and the web. flet test launches a packaged app and interacts with it on all five native platforms, and Flet says teams can now write their own pytest tests that tap buttons, enter text and verify user flows in a packaged app on desktop or mobile.

Performance work since the alpha includes a rewrite of the UI reconciliation code that Flet reports gave up to a 6.7x speed-up in control diffing in its 0.83 benchmarks. Packaged native apps now use a Python-to-Dart bridge that runs both runtimes in one process without sockets, and Android packaging was redesigned so Python packages load directly from the APK without an extraction step. The team says startup work has been reduced and bytecode compilation is on by default.

Python libraries on a phone

Flet's package index now carries more than 100 pre-built packages for iOS and Android, including NumPy, pandas, Matplotlib, Pillow, SciPy, scikit-learn, cryptography and pydantic-core. Apps can bundle Python 3.12, 3.13 or 3.14 with the app itself, and web builds use the matching Pyodide release. Availability depends on the specific package and target, and the pipeline for building wheels is called mobile-forge.

A stated compatibility policy

Flet has published a compatibility policy alongside the 1.0 release. The default deprecation period is three minor releases, with warnings and documentation pointing to replacements before an API is removed. Anything that requires action is collected in a breaking changes and deprecations page. This is the promise a Python team would want before it commits a production app to a framework: a defined window to migrate, and one place to see what breaks.

Flet 1.0 is not a drop-in replacement for the 0.28.x series. The team has published a step-by-step migration guide, and singles out the switch to a single-threaded event handler model as the change most likely to catch existing apps out.

Source

Flet 1.0 announcement on the Flet blog. Repository, star count and licence from flet-dev/flet on GitHub, verified 18 September 2026. Hacker News discussion at item 49746290.

SourceFlet

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

Share
LinkedInX
Start a project