Curator
Privacy
Last updated: 2026-09-02
Curator (discovery-curator.com) is a reading and discussion tool for working scientists. This page says what we keep about you and what happens to a lab’s Slack messages when a group imports them. It is written by the people who run the service, not by a lawyer; where a detail matters, the code is the reference and we will point you at it.
Your account
An account holds the profile you fill in (name, affiliation, the email you sign in with), the papers you save, the comments, reviews and Journal Club contributions you write, and the groups you belong to. Tokens for services you connect (ORCID, Zotero) are stored encrypted and used only to sync on your behalf. We do not sell data, run advertising, or share your reading with third parties.
To delete your account and its data, email team@discovery-curator.com from the address you signed up with. We remove the account and everything only you could see. Public comments and reviews you chose to publish are removed with it; a paper’s bibliographic record is shared catalogue data and stays. Connected services can be disconnected at any time in your settings.
Slack history import
A group manager can connect the lab’s Slack workspace to a private Curator group and import the papers the lab has discussed there. Each imported paper becomes one post in the group, backdated to when the discussion happened and labelled as imported from Slack. Here is what that involves.
What is read
Only the public channels the manager selects, through a bot the manager installs into the workspace. Nothing else: no private channels, no direct messages, no member directory. Curator does not ask Slack for permission to read names or profiles, so it never sees who anyone is.
What is stored
Short excerpts of the messages around each paper link — a thread, or the couple of hours that follow an unthreaded link — never the whole channel. Before anything is stored, user mentions, email addresses, phone numbers and links to files on Slack are replaced by placeholders. No Slack user ids and no names are kept. Each participant is represented by a pseudonym that is specific to that workspace and is used for one thing: counting how many people took part.
What it is used for
An AI model writes a short summary of the discussion for each paper — what the group made of it, the doubts and follow-ups raised — under instructions to name nobody, quote nothing and ignore anything unrelated to the paper. This is inference only: the messages are sent to the model to produce the summary and are never used to train anything. The summary is shown to the group’s members, in the group.
What leaves the group
Nothing, by default. A manager may choose to share one paper’s summary to the Hive, Curator’s public community read of that paper. There it appears as a comment clearly labelled as synthesized from a lab’s Slack discussion, attributed to the manager who shared it, and the manager can withdraw it at any time. The underlying excerpts are never shared.
How long it is kept
The excerpts stay until the manager disconnects the workspace, the group is deleted, or Slack revokes the bot’s access. Any of those deletes the excerpts. The summaries the group has already received are the group’s posts and stay until a manager removes them; a summary shared to the Hive stays until it is withdrawn.
How to object
If a post about a discussion you took part in should not be there, ask a manager of the group to remove it — every imported post can be deleted by a manager. If you would rather not go through the group, or the group has no reachable manager, email team@discovery-curator.com and we will handle it.
Contact
Questions about this page or your data? Email team@discovery-curator.com. The browser extension has its own page at /extension/privacy.