Fix watermark alpha - use gr.File instead of gr.Image to preserve PNG alpha channel (gr.Image converts to WebP yuv420p without alpha)

Fix alpha blending - use alpha=straight for proper PNG transparency handling

Fix watermark overlay - use loop filter to extend PNG duration and shortest=1 for proper alpha blending

Fix watermark alpha channel transparency - use format=yuva420p for proper alpha blending

Add watermark overlay feature - PNG with alpha channel support

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Fix: copy uploaded files to /tmp before recording to prevent Gradio cleanup

Fix iOS compatibility: H.264 profile high, level 4.2

Add pre_recording and post_recording optional video inputs

Fix: enable autoplay and force media.play() via Runtime.evaluate after load

Add: arbitrary resolution (540x960 default), audio track, post-roll (single-pass FFmpeg)

Use PNG screencast frames (lossless) instead of JPEG

Improve quality: screencast JPEG quality 100, CRF 15

Fix: wait for Page.navigate response before listening for loadEventFired

Wait for Page.loadEventFired, add duration slider (5-120s)

Hide scrollbars using --hide-scrollbars Chrome flag

Hide scrollbars via injected CSS before page render

Fix: use Emulation.setDeviceMetricsOverride for exact viewport dimensions

Fix: swiftshader WebGL, concat demuxer for correct 30s duration

Fix: ensure FFmpeg dimensions are divisible by 2 for libx264

Rewrite: direct CDP over WebSocket, no playwright, system Chromium

Add website renderer app