The Metropolitan Museum of Art (The Met) in New York publishes its collection online at a scale that covers every curatorial department: roughly 490,000 works can be searched by keyword, artist, date, medium, region and department, each with catalogue information and images. For an online visitor, the site answers a straightforward need: finding out who made a work, when, from what material, in what dimensions, where it hangs in the galleries, and what it looks like in high resolution — without setting foot in the museum. Its verifiable distinction over most museum websites is the Open Access initiative launched in February 2017: images of artworks the Museum believes to be in the public domain, together with basic data on every work in the online collection, are released under Creative Commons Zero (CC0) for any purpose, including commercial use, free of charge and without permission — backed by an open API that requires no key and a CSV dataset updated weekly.

Site overview

  • URL: https://www.metmuseum.org/ (collection entrance: https://www.metmuseum.org/art/collection)
  • Type: museum collection database and open access programme
  • Cost: free; online search and image downloads cost nothing (physical visits are ticketed separately; image licensing is covered below)
  • Registration: not required for browsing, searching or downloading
  • Interface language: site navigation is offered in English, Spanish, Portuguese, French, Italian, German, Japanese and Korean; the collection search interface and catalogue records are in English, with no Chinese version

Background

Founded in 1870, The Met operates two New York sites — The Met Fifth Avenue and The Met Cloisters — and describes its holdings as spanning "over 5,000 years of art" (Collection API documentation). The collection entrance lists 490,000+ works browseable by department (as of 2026-08-31). Digitisation of a catalogue built over more than 150 years is ongoing, and the Museum carries a content notice warning that the online collection may contain objects some viewers will find inappropriate, or information that is out of date.

On February 7, 2017 the Museum implemented its Open Access policy, making it one of the first encyclopedic museums to release public-domain images and collection data at scale. The official hub page lists Cooper Hewitt, the Art Institute of Chicago, The Cleveland Museum of Art, the National Gallery of Art and the Smithsonian Institution as peer organisations.

Searching the collection

The online collection offers keyword search with advanced filtering. Search fields include All Fields, Title, Description, Gallery, Object Number and Credit Line; filters cover Highlights, On view, Has image, Has Open Access image and Has 3D image, plus artist or culture, medium, geographic location, department and date range. A search for "van gogh", for example, returns 74 results, of which 40 carry an Open Access image (checked 2026-08-31).

The collection entrance: browse 490,000+ works by department

Public-domain works are marked with an OA icon at the lower left of the image on both result cards and object pages. Object pages add a Public Domain badge, a Download Image button, image zoom, gallery location, audio guide entries and related works; some objects, notably Egyptian and arms-and-armor pieces, offer a 3D view. Van Gogh's Wheat Field with Cypresses, for instance, is listed as on view at The Met Fifth Avenue in Gallery 822, and its high-resolution image can be downloaded directly.

Search results with the "Has Open Access image" filter applied; OA icons appear at the lower left of each card (search for "van gogh", 2026-08-31)

An object page: a Public Domain badge and Download Image button alongside gallery location

Open Access: CC0 images and data

Under the official policy, collection images fall into two categories: those of works the Museum believes to be in the public domain (or to which it waives any copyright it might have), and those it knows to be under copyright or other restrictions. Open Access covers the first category:

  • Public-domain images: the official hub states that more than 492,000 images of public-domain artworks can be downloaded, shared and remixed without restriction (as of 2026-08-31). The policy page and FAQ still cite "more than 406,000 hi-res images"; the two figures do not agree, so cite the source and date when quoting one.
  • Data for the entire online collection: basic information — title, artist, date, medium, dimensions — for every work, including those under copyright, is dedicated to the public domain under CC0, "available for any purpose, including commercial and noncommercial use, free of charge and without requiring permission".
  • Thematic sets: 20 curated image sets (Masterpiece Paintings, Cats, Vincent van Gogh, Impressionism and Post-Impressionism, and others) provide ready-made starting points.
  • Tag data: subject keywords produced by the Museum's Tagging Initiative are also CC0, in both the CSV and the API.
  • 3D and VR: more than 100 3D models of collection objects, co-produced with NHK, can be rotated and zoomed in the browser; two VR experiences cover the Temple of Dendur and Oceania highlights.

The Open Access hub describes three ways to get images and data: the online collection, the API and the GitHub CSV

The open API and dataset

The Collection API is a RESTful JSON service that requires no registration or API key, with a published rate limit of 80 requests per second. Four endpoints are available:

  • objects: all publicly accessible object IDs, filterable by data-update date and department;
  • object: the full open access record for a single work;
  • departments: the department list (19 curatorial and library departments returned on a live check, 2026-08-31);
  • search: keyword search supporting isHighlight, title, tags, isOnView, artistOrCulture, medium, hasImages, geoLocation and dateBegin/dateEnd parameters.

Object records include the isPublicDomain flag that separates open works from restricted ones, direct JPEG links to the original and smaller images, title, artist, date, medium, dimensions, accession number, credit line, subject tags, measurements (added November 2020) and Wikidata and ULAN authority links for artists and works. A live call to objects returned a total of 502,556 (2026-08-31).

The CSV dataset lives in the metmuseum/openaccess GitHub repository under CC0, is updated weekly, and is stored through Git LFS. The repository does not accept pull requests; corrections go by email to openaccess@metmuseum.org and are forwarded to the relevant department. The data is provided "as is" — cataloguing is ongoing, records vary in completeness, and the Museum advises users to refresh their copies regularly.

Copyright and content policy

The following is based on the official Open Access policy page, the Image and Data Resources FAQ and the Terms and Conditions (marked last updated October 25, 2018); all checked 2026-08-31.

What is open. Images marked with the OA icon and the basic data for the whole online collection are CC0: they can be downloaded, shared, modified and distributed for any purpose, including commercial use, free of charge and without the Museum's permission. CC0 does not affect patents, trademarks, or third-party rights of publicity or privacy. Website text is not CC0: non-OA images and written content are limited to personal, educational and fair use, and commercial publication is expressly prohibited. The CSV dataset does not include the digital images themselves; images are obtained through the website or the API.

Why some images are not open. The FAQ lists five reasons an image may lack the OA designation: the work is still under copyright or its status is unclear; privacy or publicity issues exist; the Museum does not fully own the work; contractual restrictions from the artist, donor or lender preclude open access; or the Museum has not yet produced a quality digital image. Some works have no image at all for these reasons.

Licensing restricted images. Images of works under copyright, higher-resolution versions, or new photography must be requested through the Image Request Form from Art Resource, the Museum's image-licensing partner, which charges fees by type of use. New photography is reviewed case by case, takes a minimum of six weeks, and the Museum reserves the right to deny any request. Works whose credit line reads "© ARS" require contact with the Artists Rights Society.

Attribution and trademarks. Attribution is not required for OA use, but the Museum asks that users cite it — suggested form: "The Metropolitan Museum of Art, New York" plus the credit line on the object page, optionally with www.metmuseum.org. Users may not imply the Museum's endorsement of a project or publication. Trademarks such as THE METROPOLITAN MUSEUM OF ART and THE MET are not licensed under OA and may not be used in a URL. The Museum does not issue separate license agreements for OA works.

API terms. API use falls under the same terms: the Museum logs IP addresses and usage information to monitor load; it provides no technical support and does not guarantee availability or accuracy; it may restrict access, limit transactions, require assigned keys in the future, or discontinue the service. Users are encouraged to refresh cached data, since content may be added, deleted or modified.

Use cases

  • Sourcing high-resolution public-domain art images for publication, design, course materials, exhibition collateral or data visualisation, including commercial use without license fees.
  • Art-historical research: verifying catalogue fields, accession numbers, gallery locations and credit lines.
  • Building art-related apps or data products on the API or the weekly CSV, including direct image links.
  • Trip planning: checking whether a work is on view and in which gallery.
  • Teaching with 3D models and VR content in the classroom.

Limitations

  • Works still in copyright are not open: a large share of modern and contemporary art has no CC0 high-resolution image; commercial use goes through paid licensing at Art Resource, with long review times and possible refusal.
  • Limited languages: navigation exists in eight languages, but the collection search interface and records are English-only, with no Chinese; the Japanese collection page shows a "being translated" notice (checked 2026-08-31).
  • Inconsistent figures: the number of open images differs between the hub page (492,000+) and the policy page and FAQ (406,000+ hi-res images), with no official reconciliation.
  • Uneven record completeness: cataloguing is ongoing, records vary in completeness, data is provided "as is" without accuracy guarantees, and publication-grade use requires independent verification.
  • No API service guarantees: no technical support or availability commitment, an 80 requests-per-second cap, and the Museum reserves the right to throttle, move to keys, or end the service.
  • Content notice: the Museum itself warns that the online collection may contain content inappropriate for some viewers or outdated information.

Similar sites

References