- Home
- Skills
- Workflow Automation
- Local video edit
Local video edit
Automate video editing tasks like silence removal, motion analysis, and EDL rendering using FFmpeg-powered scripts.
Free
Works with the AI Agents you already use
Local video edit
Example session with this skill installed
Find the silent parts in interview.mp4, then give me a plan to cut it down to just the speech with 0.1s padding.
- Read your context and instructions
- Compiled the local video edit
I have analyzed interview.mp4 for silence. I detected three main speech regions and generated an Edit Decision List (EDL) with a 0.1 second padding on each side to ensure no words are clipped.
Silence Detection Results
| Region | Start (s) | End (s) | Duration (s) |
|---|---|---|---|
| Speech 1 | 1.15 | 12.40 | 11.25 |
| Speech 2 | 14.20 | 28.15 | 13.95 |
| Speech 3 | 30.05 | 45.80 | 15.75 |
Proposed Edit Plan
{
"source": "interview.mp4",
"segments": [
{
"start": 1.05,
"end": 12.50,
"speed": 1.0
},
{
"start": 14.10,
"end": 28.25,
"speed": 1.0
},
{
"start": 29.95,
"end": 45.90,
"speed": 1.0
}
],
"normalize_audio": true
}
Next steps
- Review the speech regions above to confirm no critical content is missing.
- Run
python scripts/render_edl.py plan.jsonto generate the final tightened video. - Use
python scripts/normalize_loudness.pyif you need to hit a specific LUFS target for social media.
Connects securely to your tools. The creator never sees your data.
About this skill
Editing video locally is a manual, time-consuming process involving complex FFmpeg flags. This package provides a suite of Python scripts that allow an agent to analyze, trim, and master media files directly via the CLI. It handles the heavy lifting of motion analysis, silence detection, and multi-segment rendering without requiring a GUI.
What it does
- Automated Silence Removal identifies quiet gaps in dialogue and tightens the edit automatically.
- Smart Motion Analysis scores frame-to-frame changes to identify action-heavy segments or static shots.
- Edit Decision List (EDL) Rendering compiles multiple time ranges into a single finished video with optional speed adjustments.
- Audio Mastering extracts soundtracks, normalizes loudness to EBU R128 standards, and detects percussive onsets for rhythmic cutting.
- Frame Extraction captures specific timestamps or representative thumbnails based on color averages.
- Media Optimization creates H.264 proxies for faster analysis of high-resolution source files.
How it works
- Probe and Check the local environment using
check-ffmpegandprobe-mediato verify dependencies and file metadata. - Analyze the source for silence, motion, or scene cuts to generate a list of recommended keep-ranges.
- Plan the edit by presenting the detected ranges to the user for approval or modification.
- Render the final cut by passing a JSON plan to
render-edlor usingtrim-clipfor simple one-off cuts.
Frameworks & tools
This skill relies on Python and FFmpeg. It requires ffmpeg and ffprobe to be available on the system PATH.
Why this beats prompting it yourself
Writing FFmpeg commands for complex operations like filter_complex chains or two-pass EBU R128 normalization is error-prone. These scripts abstract the syntax into clean JSON inputs and outputs, ensuring frame-accurate trims and consistent audio levels every time.
Use cases
- Tightening talking-head videos by removing pauses and "um"s automatically.
- Generating highlight reels by scoring motion spikes in long recordings.
- Preparing social media clips with burned-in captions and normalized audio.
- Batch processing raw footage into smaller proxies for resource-efficient analysis.
Known limitations
Files longer than 20 minutes require a manual duration override for analysis scripts. All operations are local and require an existing FFmpeg installation on the host machine.
Requires FFmpeg and FFprobe on PATH. Files over 20m need --max-seconds. No support for transitions, overlays beyond simple text, or multi-track mixing.
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 any payment.
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 forever
No account required to browse
Trust & safety
Security scan
Security scan passed 1 day ago
- Free with an Agensi account
Requires