A plugin for AstrBot, this project addresses the need for automated group chat analysis by generating detailed reports on topics, user behavior, and key discussions. It targets users managing or interested in community dynamics across platforms like QQ, Telegram, and Discord, offering insights through LLM-powered summaries and visualizations.

How It Works

The plugin employs Large Language Models (LLMs) to extract insights such as popular topics, user profiles, and notable quotes from chat logs. It generates comprehensive reports in image and PDF formats, supporting multiple chat platforms. A key feature is "incremental analysis," which uses a sliding window to process high-volume chats efficiently, ensuring continuous coverage without context loss.

Quick Start & Requirements

Integration into the AstrBot framework is required. Key prerequisites include:

• AstrBot Framework: The base platform.
• Platform Adapters: OneBot v11 (QQ), Discord, or Telegram Bot API.
• Permissions: Discord requires Message Content Intent and View Channel History. Telegram bots not in admin roles need privacy mode disabled.
• PDF Reports: Install dependencies via /安装PDF.
• Image Generation: May require configuring external Text-to-Image (T2I) services or self-hosting. Links to T2I docs (docs.astrbot.app/others/self-host-t2i.html) and a Hugging Face service are provided.

Highlighted Details

• Multi-Platform Support: Integrates with QQ (OneBot), Telegram, and Discord.
• Rich Reporting: Generates visually appealing image and PDF reports detailing message statistics, topics, user personas, and "golden quotes."
• Incremental Analysis (Beta): A sliding window approach for continuous, efficient analysis of high-volume group chats.
• Custom LLM Integration: Allows users to specify their own LLM service providers.
• Extensible Templates: Supports custom report template contributions.

Maintenance & Community

The project acknowledges contributors and provides a DeepWiki link for information. Specific community channels (e.g., Discord, Slack) or a public roadmap are not detailed in the provided README.

Licensing & Compatibility

The project is released under the MIT License, which is permissive for commercial use and integration into closed-source projects.

Limitations & Caveats

• Experimental Multi-platform: Support beyond QQ is experimental.
• Platform Configuration: Discord and Telegram require specific permission/privacy settings.
• T2I Dependency: Image generation relies on external T2I services, potentially requiring self-hosting.
• Resource Intensive: Large message analysis can consume significant LLM tokens.
• Data Accuracy: Analysis is limited by available chat logs.
• Incremental Analysis: Currently in Beta.
• Minor Issues: Potential warnings regarding fetching quoted messages are generally non-critical.