v0.2.0-beta.11

This beta release replaces the internal system audio capture path with ScreenCaptureKit and removes the previous Core Audio Tap fallback paths.

Fixed

• Use ScreenCaptureKit as the only internal system audio capture backend.
• Fail recording startup when requested internal audio cannot start, instead of continuing as a microphone-only recording.
• Detect recordings where ScreenCaptureKit writes raw bytes for the full duration but produces only silence.
• Reject removed backend modes such as coreaudio_tap, tap, and blackhole.

Changed

• Update the native audio-capture sidecar to write ScreenCaptureKit audio as mono f32le raw PCM with metadata.
• Update internal audio documentation and availability checks to describe ScreenCaptureKit-only capture.
• Bundle the native sidecar with ScreenCaptureKit, CoreMedia, CoreGraphics, AppKit, CoreAudio, and AudioToolbox frameworks.

Tests

• uv run python tests/test_system_capture_failures.py
• uv run python tests/test_recorder_preflight.py
• uv run python tests/test_recording_pipeline_failures.py
• PYTHONPATH=. uv run python tests/test_audio_leveling.py
• uv run python tests/test_audio_devices.py
• uv run python -m py_compile src/api/recording.py src/audio/recorder.py src/audio/system_capture.py src/config.py src/api/devices.py tests/test_system_capture_failures.py tests/test_recorder_preflight.py tests/test_recording_pipeline_failures.py tests/test_audio_leveling.py tests/test_audio_devices.py
• xcrun clang ... sidecar/audio-capture/Sources/main.m -o /tmp/seam-audio-capture-test
• cargo check --manifest-path gui/src-tauri/Cargo.toml
• pnpm --dir gui build
• pnpm build
• Target PR: #75