arjncx-lang/screenme

Purpose and Runtime

Screenme is a local-only Windows 10/11 desktop screen recorder. It captures display video through FFmpeg, records microphone and system audio, and writes MP4 recordings plus optional PNG step screenshots to local user folders. The application is Python 3.10+; FFmpeg is an external executable provisioned in portable releases or by tools/download_ffmpeg.py.

Primary entry point: app/main.py. Portable bootstrap and launch are setup.bat and run_screenme.bat.

Application Composition

The application coordinates four major subsystems:

• Runtime initialization, settings, logging, platform capability detection, and FFmpeg resolution: app/main.py, app/core/settings.py, app/core/ffmpeg_finder.py.

• Capture lifecycle orchestration—start, pause/resume, stop, abort, output-session creation, and FFmpeg process control: app/capture/recorder.py, app/capture/audio_recorder.py.

• UI controls and capture-safe overlay behavior: app/ui/overlay.py, app/ui/win_exclude.py, app/ui/settings_dialog.py.

• Global keyboard interaction and tray-oriented control flow: app/core/hotkeys.py, app/main.py.

Recording Pipeline

app/capture/recorder.py is the central recording boundary. It likely translates selected capture area, frame/encoder settings, and output paths into FFmpeg screen-capture commands using Windows gdigrab, then manages the long-running encoding process and lifecycle transitions.

app/core/ffmpeg_finder.py resolves the required ffmpeg.exe; app/core/cpu_detect.py supports choosing CPU/GPU-appropriate encoding behavior. FFmpeg is the material external dependency responsible for HD video capture and MP4 encoding.

Recordings are organized as timestamped session directories containing recording.mp4, under the user’s Screenme video location. Path conventions and user storage roots are centralized in app/core/paths.py.

Audio Capture and Device Selection

Audio is separated from video capture into dedicated capture modules. app/capture/devices.py enumerates/selects recording devices, including the preference for an external microphone where available. app/capture/audio_recorder.py manages microphone and WASAPI loopback/system-audio acquisition or its integration into the recording session.

This boundary isolates Windows audio-device concerns from UI state and FFmpeg/video process orchestration.

UI, Capture Exclusion, and Region Selection

app/ui/overlay.py implements the transparent floating recorder controls and recording state presentation. app/ui/win_exclude.py applies the Windows capture-exclusion behavior (WDA_EXCLUDEFROMCAPTURE) so the control bar is not included in the recorded display.

app/ui/region_select.py owns user-driven region selection and supplies capture geometry to the recorder. app/ui/settings_dialog.py exposes persistent user choices, while app/ui/styles.py provides presentation styling rather than capture logic.

Screenshots and Session Artifacts

app/capture/screenshot.py creates PNG screenshots independently of the video lifecycle. During a recording, it emits sequential step_N.png artifacts into that active recording’s session folder; while idle, screenshots go to the user Pictures/Screenme location.

This establishes a key routing boundary: the recorder owns active-session context, while screenshot capture consumes that context when present and otherwise uses idle-output paths from app/core/paths.py.

Control Flow and Persistence

app/main.py constructs application services, initializes logging/settings, creates the overlay and settings UI, and binds UI actions and global hotkeys to recorder transitions. app/core/hotkeys.py enables Start/Stop, Pause/Resume, and Abort without requiring overlay focus.

app/core/settings.py persists capture, device, hotkey, and related user preferences across launches. app/core/logging_setup.py establishes diagnostics for startup, device discovery, FFmpeg invocation, and recording failures.