cv-mediapipe

Run MediaPipe Vision, Audio, Text, and GenAI tasks via CLI with structured JSON output.

15|4|Updated Apr 18, 2026

Search Tags:#automation #python #cli #vision #on-device #computer-vision #mediapipe

Ready to run

npx skills add https://github.com/damionrashford/media-os --skill cv-mediapipe

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: cv-mediapipe
Source: https://github.com/damionrashford/media-os/tree/main/skills/cv-mediapipe
Command: npx skills add https://github.com/damionrashford/media-os --skill cv-mediapipe

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires mediapipe, opencv-python, numpy, and includes scripts (resource) components.

What problem does it solve?

Enables running MediaPipe Vision/Audio/Text/GenAI Tasks from a single CLI without writing boilerplate Python.

Core Features & Use Cases

Supports image and video inputs for vision tasks and returns structured JSON suitable for automation.
Exposes a broad range of Tasks (vision, audio, text, genai) via .task bundles, enabling quick experimentation and integration.
Use cases include face landmark extraction on photos, object detection on video, and on-device LLM inferences.

Quick Start

Execute a MediaPipe task on an input by supplying --model and --input and optionally --mode image|video and --out-json.

Frequently Asked Questions about cv-mediapipe

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I run MediaPipe vision tasks from the command line without writing Python code?▼

Run MediaPipe vision tasks by executing a single CLI command with --model and --input arguments, returning structured JSON without needing boilerplate Python scripts.

Can I process both images and videos for object detection using MediaPipe locally?▼

MediaPipe vision tasks support both image and video inputs locally, allowing you to perform object detection and face landmark extraction while returning structured JSON output.

Do I need OpenCV and numpy installed to automate MediaPipe tasks via CLI?▼

Yes, automating MediaPipe tasks requires installing OpenCV, numpy, and mediapipe, along with supplying a valid .task bundle file to execute the operations locally.

What's the best way to extract face landmarks from photos for automation workflows?▼

Extract face landmarks from photos by running a MediaPipe vision task via CLI with a supplied .task model, returning structured JSON data suitable for integration into automation workflows.

Does this CLI approach support on-device LLM inferences and text tasks?▼

Yes, the CLI supports running on-device LLM inferences alongside audio, text, and genai tasks by utilizing MediaPipe .task bundles for local execution.

How do I get structured JSON output from MediaPipe object detection for automation?▼

Obtain structured JSON output from MediaPipe object detection by running the CLI with --out-json, which formats the vision task results for direct use in automation scripts.