VISIONModeration for game chat and what players make.
Players talk fast. Detectivision checks chat messages, player names and uploaded images for the flags your rules forbid, so your team can act before a match turns sour.
Each place, its flags
| Where | Type | Flags |
|---|---|---|
| Chat messages | text_moderation | insult, hate, racism, sexism, sexual |
| Player and clan names | text_moderation | insult, hate, sexual |
| Uploaded images and emblems | image_moderation | sexual, hate, violence |
Every account has these flags. GET /api/v1/flags lists the ones yours may send.
Name only what you need
Chat needs quick answers. Ask for the flags your rules act on: the others are not checked, so the answer comes faster.
Check names once
Check a player or clan name when it is chosen, and an emblem when it is uploaded, before anyone else sees it.
Act on reports
Moderators open the request in Logs and take one of your actions, such as Mute or Ban. Your game server gets it by webhook.
Check a chat message
curl https://api.detectivision.ai/api/v1/moderation \
-H "Api-Key: $DETECTIVISION_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"type": "text_moderation",
"input": "uninstall the game, you are trash",
"mode": "detect",
"flags": ["insult", "hate", "racism", "sexism"],
"metadata": { "playerId": "p_5521", "matchId": "r_77" }
}'{
"id": 8312,
"type": "text_moderation",
"mode": "detect",
"status": "completed",
"flagged": true,
"flags": [{ "flag": "insult", "confidence": 0.86 }],
"checked": ["hate", "insult", "racism", "sexism"],
"unchecked": [],
"tookMs": 520,
"usage": { "charged": 2, "remaining": 9998 },
"metadata": { "playerId": "p_5521", "matchId": "r_77" }
}A short message costs the same in every mode: 1 credit per started 2,500 characters by webhook, twice that at once.
Toxicity detection
Insults, hate and harassment, each a flag of its own.
Hate speech detection
Hate, racism and sexism in chat and names.
Violence detection
Violent uploaded images.
Frequently asked
Which languages are supported?
Turkish and English. Text in other languages is checked, but the results are not reliable.
How do I make chat checks faster?
Name only the flags you act on in flags. The models that feed none of them are not run, so a narrow request answers faster.
How do moderators work with it?
Every request shows in the dashboard’s Logs. You define your own actions, such as Approve or Ban; when a moderator takes one, your system gets a moderation.action webhook and carries it out. Detectivision never acts on your platform itself.
Is there a free trial?
Yes. The plans, and the free trial, are on the pricing page.