Jev-Gram

Jev-Gram is for moderating posts and comments. Keyword lists miss disguised words and also flag innocent ones, like a name that looks like a swear. Jev-Gram asks Jev and returns allow, review, or block, plus the exact spans, how severe they are, and how sure it is.

There are four strictness levels, from child-safe to anything-goes. Switching level does not call Jev again. It reapplies the scores already stored. A grey area goes to a person instead of a hard block.

It was built at the San Francisco Jev hackathon.

What you get

• JavaScript source under the MIT license
• Allow, review and block, with the words behind the verdict
• Four strictness levels from one Jev run

Claim your badge

Download the badge