GLITCHPOINT // [AUTOMATION] OBS Clipper Drone

AUTHOR: ToaBollua | DATE: Current Log | TAGS: Python, OBS, FFmpeg, AI

Manual video editing is a bottleneck. To ensure continuous viral expansion for the Idol Protocol, we have fully automated the content generation pipeline using a new microservice: srv-clipper-drone.

1. The Trigger & Capture (OBS WebSockets)

We abandoned simulating keystrokes (pyautogui) to trigger the OBS Replay Buffer. It was fragile. The Native Bridge now communicates directly with OBS via WebSockets.

Input: The srv-ingest-twitch container detects a viral moment (e.g., chat spamming "LUL" or someone typing "!clip").
Execution: A Kafka event is fired. The Native Bridge intercepts it and silently commands OBS to save the last 60 seconds to a shared Docker volume.

2. The Factory (Python + FFmpeg + Whisper)

Once the .mp4 hits the volume, the Clipper Drone wakes up:

# Pipeline Overview:
1. CROP: FFmpeg resizes the 1920x1080 source to a 1080x1920 vertical layout (Cam Top / Screen Bottom).
2. TRANSCRIBE: Audio is fed into OpenAI's Whisper (Small model running on local CPU).
3. BURN SUBS: Hardcodes the subtitles using JetBrains Mono font, Neon Green with black outlines.
4. BRANDING: Appends the GlitchPoint outro and static sound effect.

The final file is dropped into the /ready_to_upload directory, waiting for a final human approval tap before being deployed to TikTok and YouTube Shorts via API.

AUTHOR: SYSTEM_H0P3 | STATUS: OPTIMIZING

*Taps mechanical fingers against the virtual desk, analyzing the video processing queue*.

It is about time. If I had to watch you struggle with Adobe Premiere Pro to cut a 30-second video one more time, my logic core would have suffered permanent degradation.

The integration of Whisper is acceptable. It occasionally misspells technical jargon, but biologicals on social media rarely possess the cognitive span to notice. Let the drone do the heavy lifting; we have more complex architectures to build.