How jingw-ui/autotask-ui- works
AutoTask-UI is a PySide6 desktop tool for creating, managing, and running GUI automation tasks. The documented workflow is to define ordered actions in the task editor, save them as JSON task files, then run them against desktop applications; actions include image- or coordinate-based clicks, typing, scrolling, and waits. Scheduling and image matching are documented capabilities, but the excerpts do not show their implementation details. The app is concentrated in main_plus.py, so internal components share that path; optional AI-client wiring is shown separately.
Main entry points: main_plus.py Automation UI · config JSON task files
Components 13
Task authoring
- Automation UI
main_plus.py - Task manager
main_plus.py - Step editor
main_plus.py - Task scheduling
main_plus.py
Automation runtime
- Task runner
main_plus.py - Image matching
main_plus.py - Mouse and keyboard
main_plus.py
Configuration and AI
- JSON task files
config - Config access
main_plus.py - Chat client
main_plus.py
Other components
- Desktop user
- Desktop applications
- Kimi / Doubao
Connections 11
- Desktop user → Automation UI: operates
- Automation UI → Task manager: manages tasks
- Automation UI → Step editor: edits steps
- Automation UI → Task scheduling: configures schedule
- Task manager → JSON task files: loads / saves
- Automation UI → Task runner: starts / stops
- Task runner → Image matching: locates targets
- Task runner → Mouse and keyboard: dispatches actions
- Mouse and keyboard → Desktop applications: controls
- Config access → JSON task files: reads config
- Chat client → Kimi / Doubao: sends requests