[Bug] 怎么让所有的模型都走一个openai代理中转？

[q246]

📦 部署方式

Docker

📌 软件版本

当前版本：v2.15.8

💻 系统环境

Other

📌 系统版本

当前版本：v2.15.8

🌐 浏览器

Other

📌 浏览器版本

64

🐛 问题描述

能想出分隔每个api代理商的是个天才 现在几乎全部搞中转的都是仿openai的代理接口 自动分隔用其他AI官方接口全部用不了 太6
本以为直接 -e CUSTOM_MODELS="-all,gemini-2.0-flash" \ 就让其用一个模型总行吧 结果还是给我去到了谷歌的接口 为什么不直接出个openai接口可新增可减少的呢？

📷 复现步骤

No response

🚦 期望结果

No response

📝 补充信息

No response

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

Title: [Bug] How to make all models transit with an openai agent?

📦 Deployment method

Docker

📌 Software version

Current version: v2.15.8

💻 System environment

Other

📌 System version

Current version: v2.15.8

🌐 Browser

Other

📌 Browser version

64

🐛 Question description

It is a genius who can think of separating each API agent. Almost all of the people who are transiting now are imitating openai agent interfaces. Automatically separating other official AI interfaces cannot be used.
I thought it was directly -e CUSTOM_MODELS="-all,gemini-2.0-flash" \ Just let it use a model header. As a result, I went to Google's interface. Why not directly issue an openai interface that can be added or reduced?

📷 Reproduction steps

No response

🚦 Expected results

No response

📝 Supplementary information

No response

[YiJieLiuHub]

📦 部署方式

Docker

📌 软件版本

当前版本：v2.15.8

💻 系统环境

Other

📌 系统版本

当前版本：v2.15.8

🌐 浏览器

Other

📌 浏览器版本

64

🐛 问题描述

能想出分隔每个api代理商的是个天才 现在几乎全部搞中转的都是仿openai的代理接口 自动分隔用其他AI官方接口全部用不了 太6 本以为直接 -e CUSTOM_MODELS="-all,gemini-2.0-flash" \ 就让其用一个模型总行吧 结果还是给我去到了谷歌的接口 为什么不直接出个openai接口可新增可减少的呢？

📷 复现步骤

No response

🚦 期望结果

No response

📝 补充信息

No response

建议你好好看看NextChat的配置文档，里面说的很清楚了，直接gemini-2.0-flash@OpenAI就可以用了

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

📦 Deployment method

Docker

📌 Software Version

Current version: v2.15.8

💻 System environment

Other

📌 System Version

Current version: v2.15.8

🌐 Browser

Other

📌 Browser version

64

🐛 Question description

I can come up with a genius who separates each API agent. Almost all of the people who are transiting are imitating openai agent interfaces. Automatically separating other official AI interfaces cannot be used. I thought it was directly -e CUSTOM_MODELS="-all,gemini-2.0-flash" \ Just let it use a model header. As a result, I went to Google's interface. Why not directly issue an openai interface that can be added or reduced?

📷 Reproduction steps

No response

🚦 Expected results

No response

📝 Supplementary information

No response

It is recommended that you take a look at the NextChat configuration document, which is very clear. You can use it directly by gemini-2.0-flash@OpenAI.

[q246]

这个不是官方的文档吗 哪里有正版文档

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

Isn't this an official document? Where can I find the authentic document?

[Kosette]

https://github.com/ChatGPTNextWeb/NextChat?tab=readme-ov-file#custom_models-optional

有问题可以问，先搞清楚问题是什么，不要上来就火力全开。遇到脾气不好的开发者，直接把你issue关了，你又不高兴。

[YiJieLiuHub]

这个不是官方的文档吗 哪里有正版文档
写得这么明白了，不知道要怎么教。

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

Isn't this an official document? Where is the genuine document? 111111111.png
I wrote so clearly, I don’t know how to teach it.

[Oaklight]

要不然你就自己部署oneapi

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

Otherwise you can deploy oneapi by yourself