[Request] 建议请求时可以添加自定义参数

[chung1912]

🥰 需求描述

有的模型或者应用存在自定义参数，比如fastgpt的应用，cherry目前支持自定义参数，比较好用，建议lobechat可以增加这项配置，最好是可以在后端环境变量的方式，针对某个模型配置添加自定义请求参数

🧐 解决方案

增加自定义参数的配置

📝 补充信息

No response

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

🥰 Requirement description

Some models or applications have custom parameters, such as fastgpt applications. Cherry currently supports custom parameters, which is easier to use. It is recommended that lobechat adds this configuration. It is best to add custom request parameters for a model configuration in the backend environment variables.

🧐 Solution

Add custom parameters configuration

📝 Supplementary information

No response

[lobehubbot]

👀 @chung1912

Thank you for raising an issue. We will investigate into the matter and get back to you as soon as possible.
Please make sure you have given us as much context as possible.
非常感谢您提交 issue。我们会尽快调查此事，并尽快回复您。 请确保您已经提供了尽可能多的背景信息。

[arvinxx]

fastgpt的应用以及 dify 的应用在我们的设想里不应该是走 chat compeletion 接口。自定义参数这个说实话有点太 geek 了，普通用户不一定知道怎么用，我是有点存疑的。

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

In our concept, fastgpt's application and diify's application should not be the chat compelion interface. To be honest, custom parameters are a bit too geek. Ordinary users may not know how to use them. I am a little doubtful.

[chung1912]

fastgpt的应用以及 dify 的应用在我们的设想里不应该是走 chat compeletion 接口。自定义参数这个说实话有点太 geek 了，普通用户不一定知道怎么用，我是有点存疑的。

这个确实对普通用户有点难度，所以想通过后端配置，不要放在前端，不要让普通用户去配置这个。对于普通用户来说，他只要能看到这个模型名称，可以选中之后正常使用就行了

另外，如果不走chat compeletion，有什么更好的方式吗？考虑过插件或者mcp来用，但会多一个前置调用大模型，没必要

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

In our concept, fastgpt's application and diify's application should not be the chat compelion interface. To be honest, custom parameters are a bit too geek. Ordinary users may not know how to use them. I am a little doubtful.

This is indeed a bit difficult for ordinary users, so if you want to configure it through the backend, don’t place it on the front end, and don’t let ordinary users configure this. For ordinary users, as long as they can see the name of this model, they can select it and use it normally.

[arvinxx]

#4412 我们会专门做 dify 的 agent provider。这个和模型 provider 是两个

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

#4412 We will specialize in diify agent provider. This and the model provider are two