- Home
- Skills
- Image & Video
- karaoke shorts
Works with the AI Agents you already use
karaoke shorts
Automate vertical karaoke video production for YouTube Shorts and Reels with word-by-word lyric highlighting.
Free
karaoke shorts
Example session with this skill installed
Create a YouTube Short for my song 'Neon Nights'. I have neon.jpg and song.mp3. Use the lyrics: 'Driving through the city / In the pale moonlight'. Use a pink highlight for the karaoke.
- Read your context and instructions
- Compiled the karaoke shorts
I have processed your assets and generated the karaoke Short. I used htdemucs to isolate the vocals for precise timing and applied the highlight mode with color #E83454 (Neon Pink) as requested.
Video Generation Summary
| Metric | Value |
|---|---|
| Duration | 02:45 |
| Resolution | 1080x1920 (9:16) |
| Timing Accuracy | High (Vocal isolation used) |
| Special Characters | None detected |
| Output File | Neon_Nights_Short.mp4 |
Render Details
# Workflow executed:
# 1. Created source video from neon.jpg with Ken Burns zoom
# 2. Extracted vocals for timing alignment
# 3. Generated Neon_Nights.json with word-level timestamps
# 4. Built karaoke ASS subtitle file with pink highlight
# 5. Rendered final mp4 with lyrics and 3-second hook overlay
The hook title "NEON NIGHTS" was placed in the safe zone (between 13% and 87% height) to ensure it isn't cut off by the Instagram UI. I also included an equalizer animation during the 4-second instrumental intro.
Next steps
- Review
Neon_Nights.srtif you need to upload closed captions separately. - Adjust the
--delayparameter if the lyrics appear to lead or lag the audio. - Check
Neon_Nights_Short.mp4to ensure the Ken Burns zoom speed matches the song's energy.
Connects securely to your tools. The creator never sees your data.
About this skill
Producing vertical short-form content with perfectly timed lyrics is a manual bottleneck for music creators. This skill automates the generation of 9:16 YouTube Shorts and Instagram Reels from audio files, creating word-by-word karaoke overlays and visual effects without manual keyframing.
What it does
- Karaoke lyrics generates word-by-word timed subtitles with adjustable highlight or reveal modes.
- Visual generation converts a static image and mp3 into a vertical video using a Ken Burns zoom effect.
- Vocal isolation uses demucs to separate stems for high-precision timestamping.
- Dynamic overlays renders a two-line hook title and an equalizer animation during instrumental gaps.
- Shorts optimization automatically trims video to under 180 seconds with professional fades for platform compliance.
How it works
- Prepare assets by gathering the audio file and lyrics, then downloading necessary fonts like Poppins.
- Align timing using Whisper to match the audio to the text, generating a precise JSON of word timestamps.
- Generate graphics including the ASS subtitle file and the visual hook PNG.
- Final render via FFmpeg to combine audio, video, overlays, and karaoke effects into the final mp4.
Frameworks & tools
This skill uses FFmpeg for video processing and rendering, faster-whisper for speech-to-text alignment, and Pillow for title generation. It supports demucs for vocal separation and requires a Python environment.
Why this beats prompting it yourself
Writing complex FFmpeg filters for word-by-word karaoke highlighting is error-prone and time-consuming. This skill handles coordinate math for safe zones, manages font rendering for special characters, and automates the multi-step alignment process that standard LLMs cannot execute directly.
Use cases
- Convert Suno or Udio tracks into vertical videos for social media promotion.
- Create lyric videos for Turkish or multilingual songs with proper character support.
- Automate the creation of "visualizer" style reels from podcast snippets or beats.
Known limitations
Requires FFmpeg with libass support and an initial 2GB download for model weights. Rendering time depends on CPU/GPU availability for FFmpeg.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Video Skills.
Requires FFmpeg with libass support. First-run downloads for Whisper and Demucs models total ~3.6GB. Works best with high-contrast background images for lyric legibility.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 11 days ago
- Free with an Agensi account
- Software
Python
Git