Patrol MCP 是一個伺服器端套件,讓 AI 助手能管理、自動化和監控 Flutter 應用中的互動式開發會話。專為使用 Patrol 的開發者設計,此套件整合了語音輸入、語音轉文字(STT)與文字轉語音(TTS)功能,以簡化測試與除錯工作流程。伺服器可在 Linux、macOS 與 Windows 上執行,支援透過自然語言指令無縫控制 Flutter 應用的互動。它透過在開發週期中實現免手持操作來提升生產力。
適用情境
透過語音指令進行自動化 UI 測試
開發期間的免持調試
即時監控Flutter應用行為
語音控制會話自動化
與人工智慧驅動的開發助手整合
主要特色
語音驅動的 Flutter 會話控制
支援語音辨識(STT)與語音合成(TTS)整合
跨平台伺服器部署
即時互動監控
無縫整合 Patrol 生態系統
適合對象
使用 Patrol 的 Flutter 開發人員
AI輔助開發團隊
測試自動化工程師
以可存取性為重點的開發工作流程
注重效率的DevOps環境
使用建議
需要設定 Patrol 才能獲得完整功能
僅限 Flutter 開發環境
需要相容的AI助理基礎設施
需要平台特定的伺服器設定
可能需要額外的網路設定才能實現遠端存取
原始 README
以下為關聯原始碼儲存庫的 README 原文快照,最新內容請造訪原始碼儲存庫。
Patrol
A powerful, multiplatform E2E UI testing framework for Flutter apps that overcomes the limitations of integration_test by handling native interactions. Developed by LeanCode since 2022, battle-tested and shaped by production-grade experience.
Patrol is an open-source framework created and maintained by LeanCode.
However, if your company wants to scale fast and accelerate Patrol’s
adoption, we offer a set of value-added services on top of the core framework.
[!NOTE]
Patrol 4.7.0 adds Swift Package Manager support for iOS and macOS!
If you migrate your project to SPM, a few small setup changes are needed — see the iOS setup guide for details.
Patrol custom finders
Flutter's finders are powerful, but not very intuitive to use.
We took them and made something awesome.
Thanks to Patrol's custom finders, you'll take your tests from this:
Flutter's default integration_test package can't interact with the OS your
Flutter app is running on. This makes it impossible to test many critical
business features, such as:
granting runtime permissions
signing into the app which through WebView or Google Services
The CLI is needed to enable Patrol's native automation feature in integration
tests. It also makes development of integration tests much faster thanks to [Hot
Restart].
To run widget tests, you can continue to use flutter test.
See .github/WORKFLOWS.md for detailed documentation about all GitHub Actions workflows, including test schedules, Flutter versions, and deployment pipelines.