Public roadmap
What we're building next, sorted by what's actually moving. Accepted ideas come from real customer feature requests — vote on the dashboard.
Planned (1)
- PlannedModerationCivora teamFR-1001
Civora Vision — Intelligent Image and GIF Moderation
Detailed description
On modern Discord servers, communication has long ceased to be just text: members express themselves with memes, screenshots, GIFs, and various graphics. While Civora's text moderation provides outstanding security, moderating images and visual media content has required significant manual work from moderators until now. Traditional bots are completely blind in this area: they either let everything through, or block valuable, innocent images with rigid, inaccurate filters, while cunning abuses slip right past them.
The Civora Vision add-on eliminates this blind spot. This feature enables the Civora system to interpret images and GIFs sent by users with the same depth and context-sensitivity as text.
Key capabilities:
- Understanding visual context and memes: The bot doesn't just look for simple patterns; it can interpret the combined meaning of the image and the accompanying text. It recognizes hidden toxicity, harassing memes, or offensive visual elements, while intelligently letting friendly, humorous content through.
- Filtering text embedded in images: It filters out attempts where someone doesn't type rule-breaking words into the chat, but sends them in as a screenshot or graphic instead.
- Blocking graphical abuse and scams: It recognizes suspicious, deceptive visual elements in real-time – such as fake gift card graphics or phishing codes – even before they cause harm to community members.
- Optimized and customizable operation: Server administrators can fully control which channels (e.g., meme collections, media sharing channels) they want to activate visual protection on, allowing them to tailor the operation to the server's needs.
0 votes