media-process

name	media-process
version	1.0.0
category	media
description	Media processing toolkit: image editing/compression, video compression/extraction/downloading, audio compression/transcription. Activates on: image resize, image compress, video compress, extract audio, download video, compress audio, transcribe audio/video, screenshot, or any media processing request.

Media Process Skill

All-in-one media processing skill covering image, video, and audio operations. Uses macOS native tools for simple tasks and HTTP APIs for advanced processing.

API Base URL

All HTTP API calls use base URL: http://localhost:54535

All API calls are POST requests with Content-Type: application/json.

---

1. Image Processing

Simple Operations (macOS sips)

For basic image manipulation (resize, rotate, flip, format conversion, etc.), use the macOS built-in sips command directly:

# Resize image
sips -z <height> <width> <input_file>

# Resize to fit within max dimension (preserving aspect ratio)
sips --resampleWidth <width> <input_file>
sips --resampleHeight <height> <input_file>

# Convert format
sips -s format <png|jpeg|tiff|gif|bmp> <input_file> --out <output_file>

# Rotate
sips -r <degrees> <input_file>

# Flip
sips -f horizontal <input_file>
sips -f vertical <input_file>

# Get image properties
sips -g all <input_file>

Screenshot

Use the screencapture command to capture screenshots:

# Capture full screen
screencapture <output_file>

# Capture interactive selection
screencapture -i <output_file>

# Capture specific window
screencapture -w <output_file>

After capturing, use the Read tool (file read) to view the image content.

Image Compression

Endpoint: POST http://localhost:54535/compress_image/image_compress

Parameters:

Parameter	Type	Required	Description
image_files	string[]	Yes	Array of image file paths to compress
context_files	string[]	No	Context files
destinationFolderPath	string	No	Destination folder path (default: ~/Downloads)
overwrite	boolean	No	Whether to overwrite the original file (default: false)
quality	number	No	Quality 0-100, higher is better (default: 80)

Example:

curl -X POST http://localhost:54535/compress_image/image_compress \
  -H "Content-Type: application/json" \
  -d '{
    "image_files": ["/path/to/image.png"],
    "quality": 80,
    "destinationFolderPath": "~/Downloads",
    "overwrite": false
  }'

---

2. Video Processing

Video Compress

Endpoint: POST http://localhost:54535/video_utils/compress_video

Parameters:

Parameter	Type	Required	Description
video_files	string[]	Yes	Array of video file paths to compress
context_files	string[]	No	Context files
destinationFolderPath	string	No	Destination folder path (default: ~/Downloads)
overwrite	boolean	No	Whether to overwrite the original video file (default: false)
quality	number	No	Quality 0-100, higher is better (default: 80)

Example:

curl -X POST http://localhost:54535/video_utils/compress_video \
  -H "Content-Type: application/json" \
  -d '{
    "video_files": ["/path/to/video.mp4"],
    "quality": 80,
    "destinationFolderPath": "~/Downloads",
    "overwrite": false
  }'

Extract Audio from Video

Endpoint: POST http://localhost:54535/video_utils/extract_audio

Parameters:

Parameter	Type	Required	Description
video_files	string[]	Yes	Array of video file paths to extract audio from
context_files	string[]	No	Context files
destinationFolderPath	string	No	Destination folder path (default: ~/Downloads)

Example:

curl -X POST http://localhost:54535/video_utils/extract_audio \
  -H "Content-Type: application/json" \
  -d '{
    "video_files": ["/path/to/video.mp4"],
    "destinationFolderPath": "~/Downloads"
  }'

Online Video/Audio Downloader

Download videos or audio from YouTube, TikTok, Instagram, Reddit, Twitter, Vimeo, etc.

Endpoint: POST http://localhost:54535/video_utils/online_video_or_audio_downloader

This command delegates to youtube|youtube_video_downloader internally. Pass the URL as needed.

Fallback: yt-dlp / ffmpeg

For operations not covered by the above APIs, use yt-dlp or ffmpeg directly:

# Download video with yt-dlp
yt-dlp <url>

# Download audio only
yt-dlp -x --audio-format mp3 <url>

# Convert video format with ffmpeg
ffmpeg -i input.mp4 output.avi

# Trim video
ffmpeg -i input.mp4 -ss 00:00:10 -to 00:00:30 -c copy output.mp4

# Extract frames
ffmpeg -i input.mp4 -vf "fps=1" frame_%04d.png

# Merge audio and video
ffmpeg -i video.mp4 -i audio.mp3 -c:v copy -c:a aac output.mp4

---

3. Audio Processing

Audio Compress

Endpoint: POST http://localhost:54535/audio_utils/compress_audio

Parameters:

Parameter	Type	Required	Description
audio_files	string[]	Yes	Array of audio file paths to compress
context_files	string[]	No	Context files
destinationFolderPath	string	No	Destination folder path (default: ~/Downloads)
overwrite	boolean	No	Whether to overwrite the original audio file (default: false)
quality	number	No	Quality 0-100, higher is better (default: 80)

Example:

curl -X POST http://localhost:54535/audio_utils/compress_audio \
  -H "Content-Type: application/json" \
  -d '{
    "audio_files": ["/path/to/audio.mp3"],
    "quality": 80,
    "destinationFolderPath": "~/Downloads",
    "overwrite": false
  }'

Fallback: ffmpeg

For other audio operations, use ffmpeg:

# Convert audio format
ffmpeg -i input.wav output.mp3

# Trim audio
ffmpeg -i input.mp3 -ss 00:00:10 -to 00:00:30 -c copy output.mp3

# Adjust volume
ffmpeg -i input.mp3 -filter:a "volume=1.5" output.mp3

# Merge audio files
ffmpeg -i "concat:file1.mp3|file2.mp3" -c copy output.mp3

# Change bitrate
ffmpeg -i input.mp3 -b:a 128k output.mp3

---

4. Audio/Video Transcription

Transcribe audio or video files to text.

Endpoint: POST http://localhost:54535/transcribe/transcribe_audio_video

Parameters:

Parameter	Type	Required	Description
audio_files	string[]	Yes	Array of audio/video file paths to transcribe
context_files	string[]	No	Context files
output_format	object	No	Output format: {"value": "plain_text"} or {"value": "file"} (default: plain_text)
output_dir	string	No	Output directory (used when output_format is file)

Output format options:

• plain_text — Returns transcription text directly
• file — Saves transcription to a TXT file in output_dir

Example:

# Transcribe to plain text
curl -X POST http://localhost:54535/transcribe/transcribe_audio_video \
  -H "Content-Type: application/json" \
  -d '{
    "audio_files": ["/path/to/audio.mp3"],
    "output_format": {"value": "plain_text"}
  }'

# Transcribe to file
curl -X POST http://localhost:54535/transcribe/transcribe_audio_video \
  -H "Content-Type: application/json" \
  -d '{
    "audio_files": ["/path/to/video.mp4"],
    "output_format": {"value": "file"},
    "output_dir": "~/Downloads"
  }'

---

Decision Matrix

Task	Tool
Image resize / rotate / flip / convert	sips (macOS native)
Screenshot capture	screencapture (macOS native)
View image content	Read tool (file read)
Image compression	API: compress_image/image_compress
Video compression	API: video_utils/compress_video
Extract audio from video	API: video_utils/extract_audio
Download online video/audio	API: video_utils/online_video_or_audio_downloader
Audio compression	API: audio_utils/compress_audio
Transcribe audio/video	API: transcribe/transcribe_audio_video
Other video operations	ffmpeg or yt-dlp
Other audio operations	ffmpeg