About VigiMod
Moderation that reads the conversation and follows your rules
VigiMod is a content moderation API for online communities: Discord servers, forums, game and app chats. One HTTP call scores a message for harmful content and for the community's own rules, reading the messages that came before it.
Why we built it
Most moderation tools score each message on its own, with a fixed list of categories, and work best in English. Real communities are different: a joke between friends reads like an insult without the previous messages, every server has rules no generic category covers, and many chats are not in English.
Some of the tools developers relied on are going away: Google's Perspective API shuts down on 31 December 2026 and Azure Content Moderator retires on 15 March 2027. VigiMod is built for the people who need a replacement for text moderation.
Perspective API alternative →Azure Content Moderator alternative →
How it works
Context
Send up to 10 previous messages with the one to check. The score is always about the new message; the context helps to read replies, jokes and spoilers.
Your rules
Write up to 10 rules in plain language, in any language, like “No selling accounts”. Each one comes back with its own score, like a built-in category.
Scores, not verdicts
Eight categories each get a score from 0 to 1. You choose the categories, the thresholds and what to do with a flagged message.
Many languages
The model was trained on 26 languages and works best in Italian, English, Spanish, French, German and Portuguese.
What we hold to
Messages are not kept
Messages, context and rules are checked and discarded. They are never stored, logged or used to train models: only hourly usage counters per API key are kept. The only exception is a review queue you turn on for your moderators: its messages are kept for the days you choose, then deleted.
Clear about what it does
VigiMod is built for text in chats, comments and forums. We say plainly what it does not do (images, personal data) and recommend a moderator for the hardest decisions.
The community decides
VigiMod gives signals; each community sets how strict to be and what happens next. Nothing is deleted or banned by us.
Fair to everyone
We test for false flags on names, nationalities and origins: VigiMod 1.2.1 flags 0.1% of innocent messages on a held-out test, the same for every origin.
Model versions
VigiMod 1
First public model: eight categories, custom rules and conversation context.
VigiMod 1.2 · November 2026
Better custom rules and context, fewer false flags on everyday chat, child safety category opt-in.
VigiMod 1.2.1 · November 2026
Innocent messages that mention names, nationalities or origins are no longer flagged; better on threats, sexual content and context.
Who runs VigiMod
VigiMod is an independent project run by Liudmyla Vdovychenko from Italy. It follows the EU GDPR, and a data processing agreement is available to every customer. The API is free during the beta.
Contact and profiles
Questions, feedback or higher limits: write to support@vigimod.com or use the support page.
Try it on your messages
Create a free API key in two minutes, or test a message in the demo on the home page.