decord
Decord Video Loader
Decision gist · record as of 2026-08-14
Yes, if you need efficient random video frame access for deep learning and can manage system dependencies. The dormant maintenance status (no releases since June 2021, last commit July 2024) is a concern for long-term support and compatibility with newer codecs or Python versions. Install only if your video formats and Python version are confirmed compatible, and consider alternatives if active maintenance is critical for your project.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires FFMPEG/libavcodec system libraries.
- GPU acceleration (NVIDIA/Intel codecs) requires building from source with appropriate SDK.
- CPU-only wheels provided via PyPI.
License · maintenance · safety
APACHE (permissive) — Apache License (permissive) allows commercial and private use with minimal restrictions, though you must retain license notices in distributions.
last release 2021-06-14 (1887 days) · last repo commit 2024-07-17 · 2,512 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 2,437,914 downloads/mo, #3,064 on PyPI
Alternatives
Verify before relying
pip install decord
from decord import VideoReader, cpu
vr = VideoReader('video.mp4', ctx=cpu(0))
frames = vr.get_batch([0, 5, 10]) # fetch specific frames
print(frames.shape)- Whether the package works reliably with modern video codecs (H.265, VP9, AV1) beyond the FFMPEG 4.0 era mentioned in docs
- Current status of GPU acceleration support and whether NVIDIA/Intel codec implementations are production-ready
- Compatibility with Python versions beyond 3.8 (classifiers list only 3, but wheels exist for 3.6–3.8)
What it is and what it does
Decord is a video and audio decoding library designed to solve the inefficiency of random frame access during deep learning training. It wraps hardware-accelerated decoders (FFMPEG, Nvidia, Intel codecs) to provide fast, seek-optimized frame extraction from video files. The library exposes three main interfaces: VideoReader for direct frame access, VideoLoader for batched training with smart shuffling, and AudioReader for synchronized audio extraction.
The package's core strength is handling random access patterns efficiently—a common requirement during neural network training where frames are often sampled non-sequentially. It accepts numpy-like indexing, batch operations, and file-like objects for in-memory decoding. Runtime dependency is numpy only; system-level dependencies (FFMPEG, cmake, C++ compiler) are required for installation, either via prebuilt wheels or source compilation.
Use it for
- Load random frames from videos during deep learning training without seeking overhead
- Extract synchronized video and audio samples for multimodal model training
- Batch-load multiple frames at once from a single video file for efficient GPU transfer
- Shuffle and iterate over frames from many video files with optimized memory patterns
- Decode audio from video files or standalone audio files with custom sample rates
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you need efficient random video frame access for deep learning and can manage system dependencies.
The dormant maintenance status (no releases since June 2021, last commit July 2024) is a concern for long-term support and compatibility with newer codecs or Python versions. Install only if your video formats and Python version are confirmed compatible, and consider alternatives if active maintenance is critical for your project.
Install
decord on PyPI
Before you install
Medium install friction due to compiled dependencies (FFMPEG, libavcodec, cmake required). Package is dormant since June 2021 with last commit in July 2024; no active maintenance signal. Prebuilt wheels available for Linux, macOS, and Windows, but GPU acceleration requires building from source.
Requires FFMPEG/libavcodec system libraries. GPU acceleration (NVIDIA/Intel codecs) requires building from source with appropriate SDK. CPU-only wheels provided via PyPI.
License in practice
Apache License (permissive) allows commercial and private use with minimal restrictions, though you must retain license notices in distributions.
Quickstart
pip install decord
from decord import VideoReader, cpu
vr = VideoReader('video.mp4', ctx=cpu(0))
frames = vr.get_batch([0, 5, 10]) # fetch specific frames
print(frames.shape)Verify before relying
- Whether the package works reliably with modern video codecs (H.265, VP9, AV1) beyond the FFMPEG 4.0 era mentioned in docs
- Current status of GPU acceleration support and whether NVIDIA/Intel codec implementations are production-ready
- Compatibility with Python versions beyond 3.8 (classifiers list only 3, but wheels exist for 3.6–3.8)
Package facts
| License | APACHE permissive |
| Python support | Not specified |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | 1 packagenumpy |
| Maintenance | Dormant 1,887 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 2,437,914 / month, #3,064 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 3 - AlphaLicense :: OSI Approved :: Apache Software LicenseProgramming Language :: Python :: 3 |
Evidence: decord-0.6.0-cp36-cp36m-macosx_10_15_x86_64.whl; decord-0.6.0-cp37-cp37m-macosx_10_15_x86_64.whl; decord-0.6.0-cp38-cp38-macosx_10_15_x86_64.whl; decord-0.6.0-py3-none-manylinux2010_x86_64.whl; decord-0.6.0-py3-none-win_amd64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “video frame extraction”
- decordDecord decodes video and audio files with hardware-accelerated…
- torchcodecTorchCodec decodes and encodes videos, audio, and images to and from…
- scikit-videoscikit-video reads, writes, and analyzes video files using FFmpeg,…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also decord2 · torchcodec · imageio-ffmpeg · pynvvideocodec · scikit-video · nvidia-cudnn-cu12 · nvidia-cudnn-cu13 · nvidia-nvimgcodec-cu12 · Fileseq · audioread