from docs and gpt i was in the understanding that archiving is not deleting. due to an issue we decided to archive and restructure. but now it seems as it means deleting rather than archiving. we still need this data to restore our workflow and work. how can we de-archive a project or get that data available?
In the API platform site, “projects” (which contain not just API keys, but also scoped assets such as conversations, vector stores, user_data or vision uploads, etc) indeed seem to be as close to “deleted” as you can get while still having a basic listing of the prior project IDs and who archived it.
You’d have to contact OpenAI via the “help bot” at help.openai.com, ensuring login at that site first, and submitting all information about the org ID and project ID. Communicate that you need a “API organization’s project restore” feature done for you if it exists…and if all scoped data wasn’t simply discarded or made permanently inaccessible.
The red “archive” button that looks like a drawer has no hover tooltip or indication what it is for, making a press there especially dangerous even to “test” what happens.
When “archiving” by API admin endpoint, you get this guidance in the API reference:
Archives a project in the organization. Archived projects cannot be used or updated.
An archive means to me “a permanent record, a backup”, not “information gone” - you are correct that it is just one more case of bad OpenAI nomenclature.
The only thing you can change via admin endpoint is the name.
I also see fault in my own UI: I have my “default project” with the same ID in both “projects” and “archived” (with no “archived by” value).