Project title: CommonsDB: Introduction and idea generation
Brief summary: CommonsDB is a registry for public domain and freely licensed images containing a perceptual hash of them. This session will give an introduction to the project and how it can be used. It will also contain a collective idea generation exercise to find more potential benefits for the Wikimedia movement.
Expected outcomes: all participants know the basics of CommonsDB and we find more ideas on how to utilize it.
Skills preferred: no specific skills needed
Experience level: all levels welcome
Project leads: @Ainali
Starting tasks: N/A
Why are you proposing this project? Wikimedia Sverige is working on this project, and believe it can be useful for Wikimedia Commons users.
What is the expected impact? There is shared knowledge about CommonsDB and participants have their own ideas of its value.
Any other additional information for contributors: Meta overview page, some initial ideas and how they could be implemented.
Raw session notes
Note that Wikimedia Code of Conduct (https://www.mediawiki.org/wiki/Code_of_Conduct) applies in this Etherpad instance as well.
👋 Wikimedia Hackathon 2026 CommonsDB: Introduction and idea generation👋
Date & time: Friday, May 1st, 2026 at 3:15 am CEST
The use case of the CommonsDB registry is to have a place to look up info about images or other formats.
Use-case: you don't know in advance if a photo is Free or not. For example because maybe the image comes from Facebook, which wipes all metadata. So you want to know if it's a free picture.
Ainali does a live, uploading a picture from Europeana. There is a button "Generate ISCC" in Upload Wizard.
After Ainali presses the button, a sort of fingerprint is generated, and the tool is able to say that the pic comes from Europeana Foundation and it's under CC BY, etc, looking at all metadata. \o/
The tool also has an option "Copy license template to the clipboard" so you don't have to manually do that. This will be implemented during the Hackathon.
Ainali uploads another example photo in low resolution. This time the "ISCC" fingerprints is almost immediately generated. The tool is able to detect that a similar image is already available on Commons. in High Definition.
ISCC means "International Standard Content Code". The official website is:
https://iscc.codes
The tagline is "The DNA of your digital content".
The algorithm does not try to find images semantically equal. The purpose is different. It considers various things, including metadata AND pixels.
So there is one European project with 1.663.561 multimedia files already identified at the time of writing, where you can upload something:
https://registry.commonsdb.org/explorer
Ainali uploads a Monna Lisa and immediately detects it's under the public domain mark 1.0.
Ainali shows the "demo.iscc.io" where you can play with an image and see:
- the "ISCC" code
- the "DNA" visualization (a matrix of 1 and 0)
In July (in 1 month) the project wants to reach 5 millions files (now 1.6 millions).
Data stored
-local register
- image registry
- property for structured data
Participants
- at least 18 participants in person
- [[User:Valerio_Bozzolan]]
- ...
Relevant links
Phabricator task for this session: https://phabricator.wikimedia.org/T423284
Main website: https://www.commonsdb.org
Meta page: https://meta.wikimedia.org/wiki/CommonsDB
CommonsDB Registry Explorer: https://registry.commonsdb.org/explorer
Documentation: https://www.commonsdb.org/docs/
ISCC codes: https://iscc.codes
ISCC demo site: https://demo.iscc.io/#
Agenda
Basic introduction
A few examples
Ideas generation
Q&A
Notes
Integrations to existing workflows, for example, could we have an ISCC extension to OpenRefine?
A locally run app to check a folder even before upload in order to help with the selection process. Both to check for similar images within that file, but also check on the registry.
Question about other kinds of metadata, for example indigenous rights.