Introduction to AI Features

ChatNet has built-in artificial intelligence that runs across the whole platform — it can answer members from inside a chat room, moderate text and images automatically, rewrite messages, summarise conversations, translate your interface, and even write room descriptions for you. You connect one AI provider once, then switch on whichever features you want. This guide explains what the AI can do and how to turn it on.

What the AI Can Do

ChatNet ships with a single AI integration that powers several independent features. Enable as many or as few as you like.

Feature	What it does	Where it appears
AI Chat Bot	Answers members inside a room when they mention a trigger word (default `@AI`), using recent messages as context.	Chat rooms
Content Moderation	Reads each message and automatically flags or censors content that breaks the rules you choose.	Behind the scenes
Image Moderation	Scans uploaded images for adult, violent, or other unsafe content and blocks or flags them.	Behind the scenes
Message Rewrite	Lets members rephrase or clean up their own message before sending it.	Chat box
Conversation Summary	Generates a short summary of a chat conversation.	Chat
Translation	Translates your interface language files with one click.	Admin → Language
Room Description Generation	Writes a description for a chat room automatically.	Admin → Rooms

How AI Works in ChatNet

Everything is built on two ideas — providers and tasks:

Providers are the AI companies you connect with an API key. ChatNet supports OpenAI, Anthropic (Claude), Google Gemini, DeepSeek, and Grok (X.AI). You can enable more than one at the same time.
Tasks are the individual features in the table above. Each task can use your default provider, a specific provider, or be turned off — so you can point translation at a cheap model and content moderation at a more capable one.

One exception: Image Moderation does not use these chat providers. It runs on Google Cloud Vision and needs its own Google API key — see the Image Moderation guide below.

Getting Started in 3 Steps

Connect a provider. Go to General Settings > AI, pick a provider tab (for example OpenAI), set it to Enabled, paste your API key, and choose a default model. Full instructions — including where to generate each key — are in the AI Settings guide.
Assign your tasks. On the same page, the Task Configuration section lets you decide which provider handles each feature (or disable a feature entirely).
Switch on the feature pages. A few features have their own settings page you also need to enable — AI Moderation, Image Moderation, and the AI Chat Bot. The guides below walk through each one.

Detailed Guides

Guide	Use it to…
AI Settings	Connect providers, add API keys, and assign a provider to each AI task. Start here.
AI Moderation	Automatically flag or censor text messages that break your report-reason rules.
Image Moderation	Scan uploaded images with Google Cloud Vision and block or flag unsafe content.
Bot Management (AI Chat Bot)	Set up the in-room `@AI` assistant — its trigger word, personality, and message context.

A Note on Cost

AI features call your provider's API, and providers bill per use (usually per token, or per image for Vision). Short chat messages are individually cheap, but a busy community adds up. To keep costs down, assign lighter tasks such as translation or moderation to a smaller, faster model, and always monitor usage in your provider's own dashboard.