A short queue
The moderation queue is sorted by severity and groups all reports on the same content. Each case shows the content in context and offers a ready-made reason.
The institution
The AI reviews and sorts, your staff decide.
Every text is checked before it goes live, first against a banned-words list, then by the network's French AI. Reports land in a single queue, sorted by severity, with a classification suggested by the AI. Your team settles each case in one click.
What you want
“Safe moderation that takes little of our team's time.”
What Nous Campus offers
Rejected before publication
A text containing a banned word, even disguised with numbers or spaced-out letters, is not published. The member sees a message that does not repeat the word, and rephrases.
Verified listings
An offer or event is read before it appears. If nothing looks unusual, it is published right away. Otherwise the author sees the reason or a team member reviews it.
Minors come first
A report involving a minor goes to the top of the queue and the content is hidden as a precaution. Moderators receive an email alert, without the content.
Try it
Decide the cases in the queue, then write a message and watch the AI review it before it goes out.
Loading the demo…
Everything it can do
2 levelsthe banned words list, then the network's French AI
24 h or 7 daysthe two suspension durations ready to choose
What is reviewed before it goes live
- The profile: job, company, bio, background, “I like”, “I dislike”, “I offer”, “I'm looking for”, skills, links and video titles.
- Feed posts, news, requests and comments.
- Captions on photos and work, group names and descriptions.
- Offers, events, business for sale listings and pages of businesses run by members.
- The first message sent to someone you have no connection with yet.
How the review works
- A list of banned words and expressions, in French and everyday English, applied to whole words only. Words like “classic” or “assessment” go through without a problem.
- The list catches workarounds: capitals, accents, repeated letters, numbers instead of letters, letters separated by dots.
- The network's French AI catches what the list misses: disguised insults, hate speech, sexual content, harassment, threats, attacks on a person.
- For offers and events, it also checks that the listing is genuine: no soliciting, no shortened or suspicious links, no third-party contact details, no criteria prohibited by law.
- A rejected text shows the message “This text contains a word that is not accepted”, without ever repeating the word.
When someone reports
- Report in one tap on a profile, message, post, comment, offer, group, photo, event or business sale listing.
- A single queue sorted by severity: minor involved, threat, violence, harassment, sexual content, insults, then everything else.
- The AI suggests a ranking to help sort. It never decides.
- Content reported by several members, involving a minor or appearing threatening is hidden as a precaution until the team reviews it.
- Messages in group and event chat rooms also go through a filter, which sends anything suspicious to the queue.
What your team stays in control of
- One action per case: dismiss, remove, remove and warn, suspend for 24 hours or 7 days, or pause the account.
- All reported content from the same author is handled together, in a single decision.
- Every decision is justified and logged. The person receives the reasons and can appeal. The appeal is reviewed by someone else.
- Staff see every rejection, remove or restore a word on the list, add new ones, and release a false positive in one tap.
- If the AI is switched off or unavailable, the list still applies. The text is accepted and goes into the queue for review.
Your questions
Does the AI remove content on its own?
No. It rejects a text as it is being written, and the person rephrases it. For anything reported, your team decides.
Does the AI read our private messages?
It reads a private message only if it has been reported. Only the first message to a stranger goes through the banned-words list.
What if an acceptable word is rejected?
Your team sees the rejection and unblocks the text in one tap. The same text is then accepted.
Open your institution's network.
Create your account, choose your style and colours, then open your network.
Several sites or more than 30,000 members? Let's meet.