The Discourse AI plugin encompasses several modules designed to harness artificial intelligence for enhancing forum functionality. Here’s an overview of some key features:
- Embeddings Module: It automatically generates embeddings of every new post in your Discourse instance. These embeddings support semantic features like suggesting related topics and semantic topic searches, enriching the user experience by connecting them with relevant content. For more details, see the Discourse AI - Embeddings document.
- Toxicity Module: Automatically assesses the toxicity score of each new post and chat message, with the option to automatically flag content that crosses predefined thresholds. This aims to maintain community standards and user respect. Learn more in the Discourse AI - Toxicity documentation.
- NSFW (Not Safe For Work) Module: By evaluating the NSFW score of each new image upload in posts and chat messages, this module helps identify and manage content not suitable for all audiences, assisting in keeping the community safe for all users. Visit Discourse AI - NSFW for more information.
- Periodic Summary Reports Using AI: This feature within the Discourse AI / Discourse Automation plugin generates automated, high-level overviews of forum activity, providing admins and users with insights into community dynamics. Further details are provided in Introducing Periodic Summary Reports Using AI on Discourse!
- AI Bot: Offers direct integration with generative AI powered by OpenAI (ChatGPT), Anthropic (Claude), Stability AI (Stable Diffusion), and others. The AI Bot can assist with a wide range of queries, from simple questions to complex discussions. Check out how to Enable AI Bot.
For an all-encompassing view, the main Discourse AI topic provides an overview and links to various components of the plugin.
These modules represent significant steps in integrating AI into the Discourse ecosystem, aiming to enhance moderation, user engagement, and content relevance.