Datasets:

davanstrien
/

prelinger-archives-open

license: other
license_name: mixed-open
license_link: LICENSE.md
task_categories:
  - video-classification
  - text-to-video
  - visual-question-answering
language:
  - en
tags:
  - video
  - archival
  - historical
  - public-domain
  - glam
  - films
  - 20th-century
  - americana
size_categories:
  - 1K<n<10K
source_datasets:
  - original
pretty_name: Prelinger Archives Open License Videos

Prelinger Archives Open License Videos

A collection of historical films from the Prelinger Archives on the Internet Archive, filtered to include only videos with open licenses (Public Domain, CC0, CC BY, CC BY-SA).

Dataset Description

The Prelinger Archives is a collection of over 17,000 advertising, educational, industrial, and amateur films. This dataset contains the subset of videos that are available under open licenses, making them freely usable for research, machine learning, and creative projects.

What is Included

~1,900 videos with open licenses
Rich metadata including title, description, creator, date, and production details
VideoFolder format compatible with Hugging Face datasets library

Source

All videos are sourced from the Prelinger Archives collection on the Internet Archive, curated by Rick Prelinger. The collection focuses on ephemeral films: advertising, educational, industrial, and amateur films that were not intended for theatrical release.

Licenses

This dataset contains videos under the following open licenses:

License	Description
Public Domain	No known copyright restrictions
CC0	Creative Commons Zero - No Rights Reserved
CC BY	Creative Commons Attribution
CC BY-SA	Creative Commons Attribution-ShareAlike

Note: Videos with Non-Commercial (NC) restrictions are explicitly excluded from this dataset.

Each video's specific license is recorded in the licenseurl and license_type fields of the metadata.

Dataset Structure

The dataset uses VideoFolder format:

data/metadata.jsonl - Video metadata
data/*.mp4 - 1,902 video files (~143GB total)

Metadata Fields

Field	Description
`file_name`	Video filename (e.g., `video.mp4`)
`identifier`	Internet Archive identifier
`title`	Video title
`description`	Full description
`date`	Date string from archive
`year`	Year (when available)
`creator`	Creator/producer
`licenseurl`	URL to license
`license_type`	Human-readable license type
`source_url`	Link to Internet Archive page
`duration`	Video duration
`subjects`	Topic tags
`color`	Color or black and white
`sound`	Sound or silent
`country`	Country of origin
`language`	Language (when specified)
`sponsor`	Sponsoring organization
`credits`	Production credits
`production_date`	Production date
`archive_date`	Date added to Internet Archive
`collections`	Archive.org collections

Usage

Loading with Hugging Face Datasets

Note: Requires torchcodec to decode videos: pip install torchcodec

from datasets import load_dataset

# Streaming mode (recommended due to 143GB size)
ds = load_dataset("davanstrien/prelinger-archives-open", split="train", streaming=True)
for sample in ds:
    print(sample["title"])
    print(sample["description"])
    break

Content Types

The Prelinger Archives contains diverse ephemeral films including:

Educational films - Classroom and instructional content
Industrial films - Corporate training and promotional materials
Advertising - Historical commercials and promotional films
Amateur films - Home movies and independent productions
Government films - Public information and training films
Social guidance - Films about manners, health, and social behavior

Considerations

Historical Context

These films were produced primarily between the 1920s and 1980s and reflect the attitudes, language, and social norms of their time. Some content may be considered dated, offensive, or inappropriate by contemporary standards. Researchers should approach this material with appropriate historical context.

Citation

If you use this dataset, please cite:

@dataset{prelinger_archives_open,
  title = {Prelinger Archives Open License Videos},
  author = {Daniel van Strien},
  year = {2025},
  url = {https://huggingface.co/datasets/davanstrien/prelinger-archives-open},
  note = {Sourced from the Prelinger Archives on Internet Archive}
}

Please also acknowledge the Prelinger Archives and Internet Archive:

Films courtesy of the Prelinger Archives, hosted by the Internet Archive.