# facebookresearch/AugLy

A data augmentations library for audio, image, text, and video.

Repository: https://github.com/facebookresearch/AugLy
Canonical: https://ross.abutalabs.com/products/augly
Homepage: https://ai.facebook.com/blog/augly-a-new-data-augmentation-library-to-help-build-more-robust-ai-models/
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-08-24T22:44:10+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 8, longevity 100
- inputs: {"age_days": 1911, "days_push": 9, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5089, forks 311 (observed 2026-08-28T04:09:09.626958+00:00)

## What it is
AugLy is a Python data augmentation library from Meta AI supporting audio, image, text, and video with over 100 augmentations. It focuses on real-world 'internet user' transforms like memes, emoji overlays, and screenshots, useful for training robust models and evaluating robustness gaps.

## Use cases
- augment training data for image classification models
- evaluate model robustness against real-world content modifications
- generate meme-like and screenshot transforms for copy detection research
- augment text data for hate speech detection models
- apply multimodal augmentations combining text, image, audio, and video
- simulate social media reposting effects on media

## When to choose
- you need augmentations across multiple modalities in one unified API
- you want internet-platform-style transforms like emoji overlays and screenshots
- you are researching copy detection, copyright infringement, or content moderation robustness
- you need metadata about applied transforms including intensity

## When to avoid
- you only need basic standard augmentations available in your deep learning framework
- you work outside Python
- you need GPU-accelerated batch augmentation pipelines at scale
- you need a GUI-based augmentation tool

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, image-processing, audio-processing, video-processing, nlp, data-generation
- domain: machine-learning, deep-learning, computer-vision, data-science
- platform: python, cross-platform
- tags: data-augmentation, multimodal, robustness-evaluation, meta-ai, meme-transforms, copy-detection, natural-language-processing, audio, video

## Member repositories
- facebookresearch/AugLy (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:09.626958+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:02:32.365167+00:00, confidence not recorded.
  - readme: https://github.com/facebookresearch/AugLy (fetched 2026-08-28T04:09:09.626958+00:00, sha 4cb8e1b32d79)
  - homepage: https://ai.facebook.com/blog/augly-a-new-data-augmentation-library-to-help-build-more-robust-ai-models/ (fetched 2026-08-29T08:57:22.667392+00:00, sha 0f2384d14314)
  - site_page: https://ai.meta.com/about (fetched 2026-08-29T08:57:22.699605+00:00, sha 57bc8d6b3c64)
  - registry_pypi: https://pypi.org/pypi/augly/json (fetched 2026-08-29T08:57:22.752615+00:00, sha 44593767a448)
- Data as of 2026-08-30T08:39:29.467469+00:00.