AI Assistant
Entry Point
- AI assistant icon displayed at the top of the message navigation
- Tap to enter the organization default assistant conversation page
- When entering via navigation, a conversation entry is also created in the message list
- Filter AI assistant conversations in the conversation filter
Conversation Features
Multimodal Input
Supports text, images, and other input methods. The AI assistant performs deep thinking and problem analysis, replying as streaming messages.
Session Management
- Create new sessions
- View historical sessions
- Switch between AI assistant sessions
AI Skills
Conversation Message Summary
Summarizes group messages to quickly understand discussion points.
Conversation Work Summary
Generates work summaries from conversation content for review and reporting.
Web Page Reading
Provide a web page link. The AI assistant reads the content and summarizes or answers questions.
Document Reading
Intelligently reads cloud drive documents. Supports document content Q&A and summarization.
Image Text Recognition
Recognizes text in images (OCR capability).
Web Search
The AI assistant can search the web for information to answer questions requiring the latest data.
Backend Management
Model Library Management
Administrators can add two types of models:
- Public Models: General-purpose large models provided by the Lansenger platform, deployed on public cloud
- Dedicated Models: Customer privately deployed models, trained and optimized for specific data and scenarios
AI Assistant Management
- Create and manage AI assistants
- Select agent type (supports Dify Agent type)
- Add assistant basic information and skills
- Set usage scope
- Manage client entry display and custom entry icons
- Support creating third-party AI assistants

