On Wikimedia projects, anyone can include a hashtag in their edit summary when adding to the sum of human knowledge:
Learn more on the Hatnote blog.
We follow Twitter's specification (mostly) for hashtags: a hash mark (#) followed by one or more alphanumeric characters.
We do, however, ignore any hashtags which only contain numbers, as these usually denote someone counting rather than a use of a hashtag we would be interested in tracking.
You can download CSV results for a hashtag at
http://hashtags.wmflabs.org/csv/?query<hashtag>. You can also optionally provide a
project parameter to limit your results to one Wikimedia project (e.g.
enddate parameters to limit your search by date (date must be formatted as YYYY-MM-DD).
Yes! Partly. The hashtags tool was originally created by Stephen LaPorte and Mahmoud Hashemi in 2016. It was hosted on Toolforge and was used by Wikimedia editors and program leaders from all around the Wikimedia movement. By 2018, however, the tool had outgrown its setup and hosting - a database with over 8 million hashtag entries had accumulated and the tool was sluggish, often unresponsive, and was becoming a drain on other Toolforge tools' databases.
In August 2018, Sam Walton started work revitalising the tool, switching the backend from Flask to Django and setting it up on a dedicated Cloud VPS Project with dedicated database resources. This version of the tool also monitors the recentchanges MediaWiki EventStream directly instead of periodically reading from the recentchanges database table. This means it now updates in near-realtime and covers all Wikimedia projects (except Wikidata, for now). The tool no longer tracks bot-flagged edits to avoid filling up with edits that can be easily tracked in other ways (e.g. InternetArchiveBot).