- web: A simple chat interface supporting code highlighting, Latex formula rendering, and on-demand tool invocation.
- tools: Provides backend functionalities including Tokenization, content search, URL visiting, logging, etc.
- api: Access point for multiple LLM nodes, offering an OpenAI Compatible API with simple load balancing and facilitating tests across different models.
- Automatically optimize the length of dialog content
- Integrate VL series multimodal models to accept image inputs from users
- Enable reading of document formats such as PDF, DOCX, XLSX, TXT, and more.