Article written by JOHN BRANDLEY (with Andrew MacDonald)
If you create Voyant corpora for teaching, research, publications, or long-term digital humanities projects, an important change is coming to the way corpora are preserved on the voyant-tool.org server.
In brief:
Voyant corpora will no longer simply remain on the server indefinitely by default.
A new mechanism is being developed so that users can explicitly mark important corpora as
persistent and protect them from future cleanup processes.
Why is Voyant changing the way corpora are stored?
Until recently, Voyant corpora were persistent by default. There was no automatic mechanism for deleting old corpora, which meant that corpora could remain on the server for years even if they were never accessed again. Every corpus created in Voyant receives a unique corpus ID. If that ID is saved, a user can later return to Voyant, choose to open an existing corpus, and enter the ID to retrieve it.Example corpus ID:
8546b83908c4f702983ff9e55ece1492
The difficulty is that many users do not record their corpus IDs or bookmark the corresponding URLs.
Once the ID or URL is lost, the corpus may still occupy server space even though its creator can no longer
retrieve it.
Voyant also had no reliable way to determine who created a corpus, whether it was connected to a publication
or long-term project, or whether it was simply a temporary corpus that was no longer needed.
Infographic: Why saving your corpus ID mattersWhat happened when storage began to fill up?
Because every corpus was being retained, including corpora that may have been used only once, Voyant’s corpus storage eventually came close to capacity. In early 2025, a cleanup process was carried out. Corpora that had not been accessed for more than a year were deleted. This freed enough storage space for users to continue creating new corpora.The problem with the cleanup:
while it solved the immediate storage problem, some older corpora that still mattered — for example, corpora
referenced in publications or connected to long-term research — may also have been lost.
The new idea: let users mark important corpora as “persistent”
Periodic cleanup will still be necessary in the future to keep Voyant viable. However, the Voyant Consortium Board has approved a new approach developed by the Voyant Technical Council. The aim is to provide a mechanism through which a corpus owner can explicitly identify a corpus that should be protected from deletion. Such a corpus will be marked as persistent.How the persistence system is expected to work
- You create or already have a Voyant corpus.
- You keep a record of the corpus ID.
- You use your Voyant Consortium account to submit that corpus ID through the new persistence mechanism.
- Your consortium account becomes associated with the corpus as its owner.
- The marked corpus is protected from future corpus-cleanup processes.
Workflow infographic: How corpus persistence will workIs the new mechanism available now?
Not yet. The persistence mechanism is still being implemented. When it becomes available, it will be announced to the Voyant community through the newsletter. The important thing to do in the meantime is simple:If you create a corpus that you may need in the future, save its corpus ID now.
What should Voyant users do now?
- Record the corpus ID of every corpus that matters to your teaching, research, or publications.
- Keep the ID somewhere reliable, such as your research notes, project documentation, or reference manager.
- Bookmark the corpus URL where appropriate as an additional way of returning to it.
- Join the Voyant Consortium if you are not already a member.
- Watch for the newsletter announcement explaining when the persistence form becomes available.
For researchers and teachers:
This is especially important if a corpus is cited in a publication, used in a reproducible workflow,
connected to a course, or expected to support a long-term digital humanities project.