Privacy Policy
This Privacy Policy states what personal data Brilliant processes, who receives it, how long it is kept, and what you may request.
This Policy
This Privacy Policy (this “Policy”) describes how personal data is processed in connection with Brilliant. Brilliant is a Discord moderation service operated by Brilliant Bot (the “Company”). This Policy forms part of the Terms of Service. In the event of any conflict concerning liability or disputes, the Terms of Service shall control. Discord maintains its own privacy policy.
Controller and processor
The administrators of a server decide to use Brilliant and decide what is recorded. For the records of that server, those administrators are the controller and the Company is their processor. The Company is not responsible for what a server’s staff do with information.
Personal data processed
Brilliant reads messages, edits, images, links, joins, the audit log and account details in servers that add it. Most of that data is checked in memory and is not stored. The Company stores:
- moderation cases, warnings, punishments, their reasons, staff notes and active restrictions,
- AutoMod scores and the rules broken,
- Modmail conversations, which are provided to a server’s staff together with the member’s username, user ID and avatar,
- details of recent joins, short excerpts of new members’ messages, and recent bans and deletions by staff,
- server settings, suggestions, backups and licence records.
Where a server enables logging, Brilliant posts records into that server’s own channels. Cloudflare, which carries traffic to this website, and the Company’s own server receive the IP address and browser details of a visitor to this website. The bot does not collect IP addresses, email addresses, location or payment details.
Recipients
Personal data is available to a server’s staff, to the Company’s personnel where required to operate the service, to Discord, and to the processors the Company uses (MongoDB Atlas for the database and Cloudflare for the connection to this website, with the bot and the website running on our own hardware). Known scam images, scam servers, the accounts that advertised them and blocked accounts are listed across servers. A listing may cause a message to be removed, or an account to be banned, in another server.
Where a server enables Aelia, files posted in that server are read in memory, on the Company’s own machines, to detect malware. Files are not run and not stored. Aelia keeps a file’s SHA-256 fingerprint with its verdict and broad categories. Nothing about a file, its fingerprint included, is sent to any outside service.
Where Aelia removes from a member a file it finds to be malware or a message carrying a dangerous link, or stops a member it finds to be an automated account spreading spam, that member’s Discord user ID is listed in Horizon, together with the kind of removal, the number of servers in which removals occurred and the date of the last removal. Accounts listed for advertising known scam servers are included in the same way. Server owners, staff, bots, webhooks and apps are not listed by Aelia. The Company’s trust and safety team may also list an account by hand, with a category, a written reason and up to four screenshots as evidence. The reason is encrypted before it is stored. The screenshots are kept in a private channel of the Company’s staff server on Discord, each with its SHA-256 fingerprint, and the Company keeps no other copy of them. Only that team can see the reason and the screenshots. A Horizon listing does not expire. It remains until that team takes it down, and the reason and screenshots are deleted with it. Developers whom the Company has authorized may look up an account in Horizon through the Aelia API. They receive a risk level, broad categories, a range for the number of servers and the month of the last removal or listing, and never the servers, the messages, the files, the reasons or the evidence concerned.
Where the Company authorizes a developer to use the Aelia API, it keeps the developer’s Discord user ID, the teams the developer belongs to and, for each key, the application ID, privacy policy address and credit command the developer gave. For each request made with a key, it keeps the key, the route, the time, the result status and a hashed network address. It does not record what the developer looked up or the results returned.
The Company’s artificial intelligence runs on the Company’s own machines. Message text is not sent to an outside artificial-intelligence company and is not used for training. The Company’s security model is trained on simulated servers and on audit log activity from the Company’s own support server only, reduced to numbers with no names, IDs or message text. In that server the security model also reads each message once and drops the text straight away. In memory only, it keeps the time each message was sent and a one-way fingerprint of its text, for spotting copy-paste, for up to fifteen minutes. For a message it flags, it also keeps a few numbers about it and its ID for that time, and the ID for a day so the message stays out of the message logs. When a message aims abuse at another member, it keeps the time and the two members’ IDs for seven days, and never what was said, so it can tell one member who keeps going after another from friends trading insults. The Company discloses personal data where the law requires it, or to prevent harm, fraud or abuse, and personal data passes to any successor operator of Brilliant. The Company does not sell personal data or share it for advertising. Data may be processed outside your country of residence, under safeguards that applicable law accepts.
Purposes and automated decisions
The Company processes personal data to operate Brilliant, to keep records, to protect servers, to administer licences, to respond to you and to meet legal obligations. The Company relies on legitimate interests, on its agreements and on the law. It does not rely on consent. Brilliant removes messages and imposes timeouts, kicks and bans automatically, and it can be wrong. A request for review of a decision by a person shall be made to the staff of the server concerned.
Retention
The Company retains personal data for the following periods:
| Data | Retention |
|---|---|
| Modmail message text | 90 days |
| AutoMod record | 7 days after the last rule break |
| Staff monitoring | 24 hours |
| Raid monitoring | One hour after activity ceases |
| Horizon listing made by Aelia | Until the Company takes the listing down |
| Horizon listing made by staff, with its reason and evidence | Until the Company takes the listing down |
| Aelia API request record | 30 days |
| All other personal data | Until the server deletes it, its owner requests deletion, or the Company removes it |
Removal of Brilliant from a server does not delete that server’s records. The Company may retain personal data for a longer period where applicable law permits.
Your rights
A request for a copy of your personal data, or for a correction, shall be sent to [email protected] and shall include your Discord user ID. Moderation records are not deleted at a member’s request. Those records belong to the server that made them. Only the owner of a server may request deletion of that server’s data.
Depending on your place of residence, you may have further rights, including a right to object, a right to human review of an automated decision, and a right to complain to a data-protection authority. The Company verifies requests, responds within the time applicable law allows, and may refuse a request to the extent the law permits refusal. Privacy rights that applicable law does not permit to be waived are unaffected.
Children, security and changes
Brilliant is not directed to children under the age of thirteen (13). If the Company is notified at [email protected] that a person is under thirteen (13), it shall delete the personal data it holds and can identify as that person’s, apart from a record of the request. A person must be old enough to use Discord. A person under eighteen (18), or under the age of majority where that person lives, may use Brilliant only with the permission of a parent or legal guardian, as stated in the Terms of Service. Sensitive text is encrypted. No system is perfectly secure, and the Company does not guarantee that personal data will never be accessed, lost or altered. This website sets no cookies of its own and uses no analytics or tracking. Cloudflare may set a security cookie. The Company may amend this Policy at any time by posting a revised version on this page. Any claim concerning this Policy, or concerning personal data Brilliant processes, is subject to the release, the limitation of liability, and the notice, time-limit, arbitration and individual-claim provisions of the Terms of Service.