Powered by GPT-4o & Whisper

Turn Long Videos Into
Viral Short-Form Clips

ClipMind analyzes your long-form content with AI, detects the most compelling moments, and automatically renders them into 9:16 clips — complete with dynamic captions, face tracking, and social media copy.

Get Started on GitHub See How It Works
📹 Video Timeline — Viral Segments Detected
🔥 97
🤯 91
💡 88
😂 95
4
Clips Found
97
Top Score
9:16
Auto-Reframe
1080p
Export Ready
GPT-4o
Viral Analysis Engine
9:16
Auto Face-Tracked Reframe
3+
Platforms Supported
1080p
Export Quality

Everything You Need to
Go Viral — Automatically

A complete AI-powered pipeline that takes raw footage and delivers platform-ready short-form clips.

🎯
AI Viral Detection
GPT-4o scans your transcript and identifies 3–5 high-potential moments with viral scores, hook text, and strategic timestamps.
🎙️
Whisper Transcription
Word-level Whisper timestamps drive every caption, SFX trigger, and edit point. Results are cached so you never pay twice.
📱
9:16 Auto-Reframe
MediaPipe face detection with smoothed OpenCV tracking keeps the speaker centered in portrait mode — automatically.
✍️
Dynamic Captions
Pillow-rendered 2–4 word captions with per-word active highlighting in gold. Full Arabic and English support with RAQM shaping.
📣
Social Media Hub
Generates 3 alternative titles, SEO descriptions, CTAs, and hashtags for YouTube, TikTok, and Facebook in one click.
🔊
Word-Synced SFX
Optional pop/whoosh sound effects fire in sync with caption words for maximum engagement on TikTok-style content.
🎞️
EDL Export
Writes CMX 3600 Edit Decision List files for seamless import into Premiere Pro or DaVinci Resolve workflows.
🪄
SRT Alignment
Optionally supply corrected .srt subtitles as ground truth. ClipMind fuses SRT text with Whisper word timestamps for perfect Arabic captions.
🖥️
Dark-Mode Desktop GUI
PySide6 drag-and-drop interface with viral score cards, caption style controls, threaded processing, and live logs — no terminal required.

From Raw Footage to
Ready-to-Post in Minutes

A fully automated pipeline — drop a video, get viral clips.

1
Drop Your Video
Drag any MP4, MOV, MKV, or AVI file into the GUI or pass it via CLI.
2
Whisper Transcribes
FFmpeg extracts audio. Whisper generates word-level timestamps. Results cached locally.
3
GPT-4o Analyzes
The AI scans for hooks, emotion, conflict, and punchlines — returning strict JSON clip metadata.
4
Clips Are Rendered
FFmpeg cuts segments. MediaPipe reframes to 9:16 with face tracking. Captions and SFX are burned in.
5
Post & Profit
Social Media Hub generates titles, descriptions, CTAs, and hashtags per platform. One-click copy.

Built With Best-in-Class
AI & Media Tools

A carefully chosen stack of state-of-the-art AI models and battle-tested media processing libraries.

OpenAI GPT-4o
Whisper (word-level)
FFmpeg & FFprobe
OpenCV
MediaPipe
Pillow (RAQM)
PySide6 / Qt
Pydantic
Python 3.10+
CMX 3600 EDL
PowerShell — ClipMind CLI
1# Launch the desktop GUI
2clipmind gui
3 
4# Full pipeline — detect & render 5 viral clips
5clipmind process input_videos/long_video.mp4 --clips 5
6 
7# With reference SRT for perfect Arabic captions
8clipmind process video.mp4 --srt video.srt --clips 5
9 
10# Dry-run: analyze only, no rendering (saves cost)
11clipmind process video.mp4 --dry-run

Ready to Make Your
Content Go Viral?

Clone the repo, add your OpenAI key, drop a video. That's all it takes.