wafle-video-processor

Automatic watermark removal from videos via FFmpeg delogo filter.

Purpose: Remove Meta AI / Facebook watermarks from downloaded videos. Target: Semi-transparent "Meta AI" logo, static bottom-right position. Requirement: No GPU — runs on CPU, faster than real-time.

Install

pip install wafle-video-processor

Requires FFmpeg in PATH:

winget install FFmpeg
# or: https://ffmpeg.org/download.html

CLI Usage

# Single file (in-place)
wafle-video-processor remove video.mp4

# Single file (new file)
wafle-video-processor remove video.mp4 -o clean.mp4

# Batch directory
wafle-video-processor batch ./videos/ -o ./cleaned/

# Custom watermark position/size
wafle-video-processor remove video.mp4 --position bottom-right --width-pct 8 --height-pct 4

Python API

from waflevideoprocessor import remove_watermark, has_ffmpeg

if not has_ffmpeg():
    print("Install FFmpeg first")

result = remove_watermark("video.mp4", output_path="clean.mp4")
print(result["success"])  # True/False

Integration

Automatically called by wafle-automations-meta after each video download when wafle-video-processor is installed.

Meta AI Watermark

• Semi-transparent "Meta AI" logo
• Bottom-right corner
• ~10% width × ~5% height
• Static position across all free-tier videos