Does it work on macOS hardware?

Yes, it includes specific optimizations for Apple Silicon and Intel Macs using the VideoToolbox framework for both H.264 and HEVC codecs.

Does this skill support the latest FFmpeg 8.0 features?

Yes, it includes documentation for FFmpeg 8.0.1 features released in late 2024 and 2025, such as pad_cuda and updated Vulkan compute codecs.

Can I use this for AV1 encoding?

Absolutely. The skill provides configurations for av1_nvenc (NVIDIA RTX 40 series and newer), av1_qsv, and av1_amf.

What is the 'Full GPU Pipeline' mentioned in the documentation?

It is a method that keeps video frames entirely within GPU memory during decoding, filtering, and encoding to avoid the performance bottleneck of transferring data over the PCIe bus to the CPU.

How much faster is hardware acceleration compared to CPU encoding?

GPU encoding typically ranges from 10x to 30x faster than software-based CPU encoding, though it may trade off a small amount of compression efficiency for this massive speed increase.

FFmpeg Hardware Acceleration

Name: FFmpeg Hardware Acceleration
Author: JosiahSiegel

byJosiahSiegel

•

Developer Tools

Accelerates video encoding and decoding using hardware-specific GPU pipelines for NVIDIA, Intel, AMD, and Apple systems.

About

This skill provides a comprehensive toolkit for implementing high-performance video processing using FFmpeg 7.1 and 8.0. It offers specialized commands for NVIDIA NVENC/NVDEC, Intel Quick Sync (QSV), AMD AMF, and Apple VideoToolbox, along with Linux VAAPI and Vulkan Video 8.0 support. By leveraging dedicated hardware engines, it enables video transcoding speeds up to 30x faster than CPU-only methods. It includes advanced patterns for full GPU pipelines, including CUDA-accelerated scaling, padding, and filtering to minimize PCIe transfer bottlenecks between CPU and GPU memory.

Key Features

7 GitHub stars
Vulkan Video support for modern codecs including AV1, VP9, and ProRes RAW
Advanced quality tuning with CQP, VBR, and lookahead optimization strategies
Multi-platform support for NVIDIA NVENC, Intel QSV, AMD AMF, and Apple VideoToolbox
Optimized GPU pipelines using scale_cuda, pad_cuda, and overlay_cuda for FFmpeg 8.0+
Hybrid CPU/GPU filtering guides using specialized hwupload and hwdownload patterns

Use Cases

High-speed batch video transcoding for media servers, CDNs, or large-scale archival
Real-time video streaming and live encoding with low-latency hardware presets
GPU-accelerated video editing pipelines within Docker containers or cloud environments

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add josiahsiegel/claude-plugin-marketplace ffmpeg-hardware-acceleration

For use in Claude.ai and ChatGPT

Download Skill

GitHub