- Any chat added through the upper right corner will carry all historical messages that have been successfully sent
- Voice messages currently only support ChatGPT’s Whisper model, and voice conversations only support ChatGPT’s Whisper and TTS models
- The purpose of more functions in the upper right corner of the voice chat page: The principle of voice chat is to first convert your voice content into text through the Whisper model, then call the text model to generate content, and then play the content through TTS , and the model selected in the More button is the function of calling the text model to generate content. The advantage is that you can choose any model to generate content instead of the fixed ChatGPT model, which allows for more flexible content generation.
- Anywhere where a model version is not selected will be set as the default model version on the service provider list page
- The service page supports adding ChatGPT and Gemini official and third-party API Keys and Servers
- Click the plus sign in the upper right corner to enter the new page, enter the alias API Key and API Server, and click Save. When saving, all models supported by the API Key will be obtained. After successful saving, one model will be defaulted as the current model. , which can be modified on the service page
- On the service list page, after entering the specific service provider list, click on each piece of data to set the piece of data as the default model. Click on the model version to the right of the name to switch to the default model. The default model only supports text content. Model switching
- The default chat page features Casual Chat and Generate Pictures. After entering Casual Chat, you will find the default model and corresponding model version set in the service. If your model supports Whisper, there will also be a voice input option. Click the title bar to switch text model versions. This casual chat will not carry any contextual information and is suitable for casual questions
- Generating pictures supports generating pictures. When generating pictures, they will be generated according to the default model and model version set in the service. The generated pictures can be saved to the album. Click on the picture to view the larger picture. Click on the title bar to switch the text model version.
- Click the plus sign in the right corner to add customized chat content. Note that the chat will carry all contextual messages, which is suitable for long content chats. If you do not want to carry the context, you can add a custom chat again, or click Feel free to chat at the top of the page
- Requires a model that supports TTS and Whisper to use.
- The settings page supports setting whether to generate titles. Generating titles may cause the loss of some tokens. You can set it according to your own needs.
- Support switching UI, night mode, following system, etc.
- 任何通过右上角新增的聊天，聊天时都会携带所有已经发送成功的历史消息
- 语音消息目前只支持ChatGPT的Whisper模型，语音对话也只支持ChatGPT的Whisper和TTS模型
- 语音聊天页面，右上角的更多功能的目的：语音聊天的原理是先将你的语音内容通过Whisper模型转换成文字，然后调用文字模型生成内容，然后再将内容通过TTS播放出来，而更多按钮里选择的模型是调用文字模型生成内容这一步的作用，他的好处是你可以选择任意的模型来生成内容，而不是固定的ChatGPT的模型，这样可以更加灵活的生成内容。
- 所有没有选择模型版本的地方，都会设置为服务商列表页面的默认模型版本
- 服务页面支持添加ChatGPT和Gemini官方和三方的API Key和Server
- 点击右上角的加号进入新增页面，输入别名API Key和API Server，点击保存即可，保存的时候会获取API Key的支持的所有模型，保存成功之后，会默认一个模型为当前模型，可以在服务页面进行修改
- 在服务列表页面，进入特定的服务商列表后，点击每一条数据，会将该条数据设置为默认模型，点击名称右侧的模型版本，可以切换默认模型，默认模型仅支持文本内容的模型切换
- 聊天页面默认置顶的有随便聊聊和生成图片，随便聊聊进入后会查找服务中设置的默认模型和对应的模型版本，如果你的模型支持Whisper，还会有语音输入的选项。点击标题栏可以切换文本模型版本。 该随便聊聊不会携带任何上下文消息，适合随问随走
- 生成图片支持生成图片，生成图片的时候会根据服务中设置的默认模型和模型版本进行生成，生成的图片可以保存到相册，点击图片可以查看大图，点击标题栏可以切换文本模型版本。
- 点击右商角的加号可以添加自定义的聊天内容，注意，该聊天会携带所有的上下文消息，适合长内容聊天，如果不想携带上下文，可以再次新增自定义聊天，或者点击置顶的随便聊聊
- 需要支持TTS和Whisper的模型才能使用。
- 设置页面支持设置是否生成标题，生成标题可能会有部分token的损耗，可以根据自己的需求进行设置
- 支持切换UI，夜间模式，跟随系统等等