Chat Interface

1. Assistants and Topics

a. Assistants

Assistant Configuration
The assistant configuration function provides preset general assistants and customizable agents, supporting flexible adjustment of configuration items such as preset prompts and model parameters to build exclusive conversation assistants. As the core configuration unit, its parameter settings will be automatically applied to all subordinate conversation instances, ensuring the consistency of the conversation environment.

Topic Management
Topic management supports creating independent conversation threads under a specified assistant environment, realizing simultaneous management and maintenance of multiple topics. This function provides topic name modification and content cleaning capabilities, and ensures that all topics automatically follow the assistant's preset complete parameter configuration, maintaining the uniformity of the conversation environment configuration.

b. Chat

Chat covers three core capabilities: conversation environment management, multi-modal content interaction, and external resource integration. The system supports the creation and maintenance of independent conversation sessions, features real-time content clearing and context reset functions, and can parse image and document content and provide instant translation services. Through online retrieval, knowledge base calls, and MCP service docking, external resource integration is achieved, while providing configuration options such as dynamic model switching, quick call of preset instructions, and adaptive interface extension, building an efficient human-computer interaction experience.

2. Chat Settings

a. Model Settings

This setting inherits and synchronizes model parameters from the assistant configuration. In the current conversation environment, only the model setting items are effective for the current assistant, and other configurations are applied globally.

b. Message Settings

The message settings module provides complete interactive interface customization capabilities, supporting visual editing of prompts and format specification verification, serif font and CSS style switching, automatic folding of thinking processes, structured indexing of conversation content, multi-model answers, and flexible adjustment of message font size, comprehensively optimizing information presentation efficiency and visual experience.

c. Input Settings

The input optimization module integrates real-time Token monitoring, long text file conversion, independent Markdown rendering control for user and model messages, triple-click space quick translation mechanism, and flexible configuration of multi-language translation targets, comprehensively improving input interaction efficiency and content processing accuracy.

3. Assistant Settings

a. Edit Assistant

Assistant settings allow users to fine-tune the model behavior of the assistant to meet the needs of different scenarios. Assistant settings mainly include model settings, model temperature, context count, streaming output, and maximum token count. Assistant settings apply to all topics under that assistant.

b. Model Settings

Model settings provide a complete conversation environment configuration solution, supporting setting an exclusive default model for the assistant (priority higher than global settings), and featuring the maintenance function of automatically restoring the preset model when creating a new topic. Context memory capacity can be flexibly adjusted to adapt to task requirements of different complexities, and streaming transmission and whole segment return modes are provided to ensure the best balance between conversation coherence and real-time performance.