Overview & result
A Discord moderation bot written in Python on TypeSafe's Jev. Per the README it checks messages as they arrive, deletes spam and scam links with a reason in the server audit log, and escalates repeat offences from warning DMs to timeouts. A /profile command scores a member's recent messages for scam risk, spamminess, toxicity and helpfulness.
From the original post
A Discord moderation bot built with Python and TypeSafe AI (Jev System One). It filters spam and scam links in real time, escalates offenses automatically, and lets moderators profile members based on their message history.
How Jev fits in the loop
- The bot checks each Discord message as it arrives.
- Messages flagged as spam or scam links are deleted, with the reason written to the server audit log.
- Repeat offences escalate from a warning DM to a final warning, a 10-minute timeout, then 1-hour timeouts.
- When an admin pardons a false flag, the message is stored and fed into later evaluations as a legitimate example.
How to reproduce
- Inspect the original showcase and source material at https://github.com/brainstormity/Jev-Moderation-Bot
- Verify the bounded prompt and input candidate schema configured for Jev
- Benchmark decision latency and classification accuracy against baseline models
Why this build matters
The false-flag memory lets a server teach the bot its own slang, links and jokes without retraining.
Limitations
- The README gives no accuracy or false-positive figures.
- User profiling reads only cached messages or the history of a channel you pick.
- Timeouts are configurable defaults, and pardons depend on moderators reviewing the mod log.