1.10.4

• fix: Missing items on the page when selecting the model
• fix: dell-e 3 default image size
• feat: Plugin supports direct output of binary images

1.10.3

• fix: Duplicate saving of user input in historical messages bug
• feat: Support for image input in document format

1.10.0

feat: switching Models with InlineKeyboards

1.9.4

• fix: bool type conversion error bug in configuration #290

1.9.3

• feat: add a plug-in system
• fix get/set USER_CONFIG bug
• fix send workers image bug

1.8.0

• Support Cohere, Anthropic Ai
• Support image input.
• Adapt to group topic mode
• Remove the role function and use custom commands instead.
• Fix the bug of failure to send super long text.

1.7.0

• Modify the worker AI invocation method to API invocation, requiring the setting of account_id and token. The original AI binding method is invalid.
• Add support for worker AI text conversation flow mode.
• Add functionality for worker AI to generate images from text.
• Add switch AI providers
• Add custom commands, which allows for quick model switching
• Add lock user-defined configurations

v1.6.0

• Add workers AI support, please refer to the configuration document for specific settings.
• Optimize the parser for openai stream mode.

v1.5.1

• fix 1.5.0已发现BUG

v1.5.0

• perf: 调整命令顺序
• perf: openai发送请求前前发送loading消息
• feat: 添加流式输出支持(默认开启)。使用STREAM_MODE=false关闭
• feat: 添加安全模式(默认开启)解决TG无限重试的问题
• feat: 增加对多个KEY的适配，随机选择KEY使用
• feat: 增加快捷按钮 /new, /redo