NextChat [Feature Request] 请求新增对火山引擎DeepSeek R1联网版api的支持

🥰 需求描述

目前火山引擎部署的普通版DeepSeek-R1是可以正常调用的，比如使用 url=https://ark.cn-beijing.volces.com/api/v3, model=ep-20xxx217xxxx26-rxxxx。但是火山引擎支持DeepSeek-R1的联网版，并支持api调用。调用示例如图下，Chat Box 已支持使用，是否可以新增对这种api调用的支持？这样就可以无需插件，自动联网搜索。

🧐 解决方案

新增对火山引擎部署的DeepSeek-R1的联网版的支持。url=https://ark.cn-beijing.volces.com/api/v3/bots (和普通版的相比加了一个/bots后缀）， model=bot-xxxxx21721xx02-xx5wp

📝 补充信息

No response

Feb 18 '25 02:02 WadeQAQ

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

Title: [Feature Request] Request new support for the volcanic engine DeepSeek R1 networked version API

Feb 18 '25 02:02 Issues-translate-bot

您好请问您能分享一下网站上的接口地址模型名是怎么配置的吗我用的火山方舟的DeepSeek-R1的api 但是一直没有办法对接上,我的配置如下：

错误如下：

{
  "error": {
    "message": "Incorrect API key provided: 6d96da5a************************1e5e. You can find your API key at https://platform.openai.com/account/api-keys.",
    "type": "invalid_request_error",
    "param": null,
    "code": "invalid_api_key"
  }
}

谢谢

Feb 18 '25 16:02 yangnan996

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

Hello, can you share the interface address on the website. How is the model name configured? I use the DeepSeek-R1 API of Volcano Ark, but it has not been able to connect to it. My configuration is as follows:

The error is as follows:

{
  "error": {
    "message": "Incorrect API key provided: 6d96da5a***************************1e5e. You can find your API key at https://platform. openai.com/account/api-keys.",
    "type": "invalid_request_error",
    "param": null,
    "code": "invalid_api_key"
  }
}

Thanks

Feb 18 '25 16:02 Issues-translate-bot

您好请问您能分享一下网站上的接口地址模型名是怎么配置的吗我用的火山方舟的DeepSeek-R1的api 但是一直没有办法对接上,我的配置如下：

错误如下：

{ "error": { "message": "Incorrect API key provided: 6d96da5a************************1e5e. You can find your API key at https://platform.openai.com/account/api-keys.", "type": "invalid_request_error", "param": null, "code": "invalid_api_key" } } 谢谢

如果只是调用不带联网功能的字节跳动的api的话, 我的环境变量中设置的是: "
CUSTOM_MODELS=+deepseek-v3@ByteDance=ep-202xxx1720xxxx-rxxxl BYTEDANCE_API_KEY=xxxxxxxxxxxxxxxxx "

Feb 19 '25 01:02 WadeQAQ

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

Hello, can you share the interface address on the website. How is the model name configured? I use the DeepSeek-R1 API of Volcano Ark, but I have not been able to connect to it. My configuration is as follows:

The error is as follows:

{ "error": { "message": "Incorrect API key provided: 6d96da5a***************************1e5e. You can find your API key at https://platform .openai.com/account/api-keys.", "type": "invalid_request_error", "param": null, "code": "invalid_api_key" } } Thanks

If you just call ByteDance's API without networking, the settings in my environment variable are: " CUSTOM_MODELS=+deepseek-v3@ByteDance=ep-202xxx1720xxxx-rxxxl BYTEDANCE_API_KEY=xxxxxxxxxxxxxxxxxxxxxxxxxxxxx "

Feb 19 '25 01:02 Issues-translate-bot

Custom_models =+deepseek-r1@bydance = ep-202xxx1720xxxxxxxxl Bydance_url = https: //ark.cn-beijing.volces.com/api/v3/ 这样写提示 Empty Response from Server 是不支持吗

Feb 20 '25 12:02 NUAAwanghe

同求，火山联网版支持，自定义加不上啊，在cherry studio上是可以通过兼容openai API添加

Feb 20 '25 15:02 jfc234

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

As for the same question, Volcano Network version supports it, customization cannot be added. It can be added through the compatible openai API on cherry studio.

Feb 20 '25 15:02 Issues-translate-bot

Custom_models =+deepseek-r1@bydance = ep-202xxx1720xxxxxxxxl Bydance_url = https: //ark.cn-beijing.volces.com/api/v3/ 这样写提示 Empty Response from Server 是不支持吗已经解决，要这样设置BYTEDANCE_URL=https: //ark.cn-beijing.volces.com/ CUSTOM_MODELS=+deepseek-r1@ByteDance=ep-202xxx1720xxxxxxxxl,+deepseek-v3@ByteDance=ep-202xxx1720xxxxxxxxx 注意服务器设置dns 114.114.114.114

Feb 21 '25 01:02 NUAAwanghe

想知道大佬是怎么配置的，我的接口地址用的https://ark.cn-beijing.volces.com/api/v3，但是显示 { "error": true, "message": "empty response from server" } 还有模型ID是用火山引擎在线推理发的那个id吗（明明我直接在python里用openai的SDK，然后这样配置是可以正常使用的

Feb 21 '25 09:02 txsxuan

想知道大佬是怎么配置的，我的接口地址用的https://ark.cn-beijing.volces.com/api/v3，但是显示 { "error": true, "message": "empty response from server" } 还有模型ID是用火山引擎在线推理发的那个id吗（明明我直接在python里用openai的SDK，然后这样配置是可以正常使用的

把 /api/v3去掉模型ID就是创建模型下面有个ep开头的模型ID 添加模型这样写 +deepseek-v3@ByteDance=ep-202xxx1720xxxxxxxxx

Feb 21 '25 10:02 NUAAwanghe

想知道大佬是怎么配置的，我的接口地址用的https://ark.cn-beijing.volces.com/api/v3，但是显示 { "error": true, "message": "empty response from server" } 还有模型ID是用火山引擎在线推理发的那个id吗（明明我直接在python里用openai的SDK，然后这样配置是可以正常使用的

把 /api/v3去掉模型ID就是创建模型下面有个ep开头的模型ID 添加模型这样写 +deepseek-v3@ByteDance=ep-202xxx1720xxxxxxxxx

参照你的意见，这样启动 docker run -d -p 3000:3000
-e BYTEDANCE_API_KEY=8XXXXXXX80-c8cd-465XXXXXXXX
-e BYTEDANCE_URL=https://ark.cn-beijing.volces.com
-e CUSTOM_MODELS=-all,+DeepSeekV3@ByteDance=ep-2025XXXXXX-jXXXXXXXpc
-e CODE=XXXXXXX
yidadaa/chatgpt-next-web

在网页对话后还是响应以下错误 { "error": true, "message": "Failed to fetch" } { "error": true, "message": "empty response from server" }

求问大佬怎么启动的docker，求完整启动命令，感谢

Feb 22 '25 06:02 Week888

想知道大佬是怎么配置的，我的接口地址用的https://ark.cn-beijing.volces.com/api/v3，但是显示 { "error": true, "message": "empty response from server" } 还有模型ID是用火山引擎在线推理发的那个id吗（明明我直接在python里用openai的SDK，然后这样配置是可以正常使用的

把 /api/v3去掉模型ID就是创建模型下面有个ep开头的模型ID 添加模型这样写 +deepseek-v3@ByteDance=ep-202xxx1720xxxxxxxxx

参照你的意见，这样启动 docker run -d -p 3000:3000 -e BYTEDANCE_API_KEY=8XXXXXXX80-c8cd-465XXXXXXXX -e BYTEDANCE_URL=https://ark.cn-beijing.volces.com -e CUSTOM_MODELS=-all,+DeepSeekV3@ByteDance=ep-2025XXXXXX-jXXXXXXXpc -e CODE=XXXXXXX yidadaa/chatgpt-next-web

在网页对话后还是响应以下错误 { "error": true, "message": "Failed to fetch" } { "error": true, "message": "empty response from server" }

求问大佬怎么启动的docker，求完整启动命令，感谢

你仔细看清楚模型写法，DNS是否配置能访问https://ark.cn-beijing.volces.com

Feb 22 '25 06:02 NUAAwanghe

想知道大佬是怎么配置的，我的接口地址用的https://ark.cn-beijing.volces.com/api/v3，但是显示 { "error": true, "message": "empty response from server" } 还有模型ID是用火山引擎在线推理发的那个id吗（明明我直接在python里用openai的SDK，然后这样配置是可以正常使用的

把 /api/v3去掉模型ID就是创建模型下面有个ep开头的模型ID 添加模型这样写 +deepseek-v3@ByteDance=ep-202xxx1720xxxxxxxxx

参照你的意见，这样启动 docker run -d -p 3000:3000 -e BYTEDANCE_API_KEY=8XXXXXXX80-c8cd-465XXXXXXXX -e BYTEDANCE_URL=https://ark.cn-beijing.volces.com -e CUSTOM_MODELS=-all,+DeepSeekV3@ByteDance=ep-2025XXXXXX-jXXXXXXXpc -e CODE=XXXXXXX yidadaa/chatgpt-next-web 在网页对话后还是响应以下错误 { "error": true, "message": "Failed to fetch" } { "error": true, "message": "empty response from server" } 求问大佬怎么启动的docker，求完整启动命令，感谢

你仔细看清楚模型写法，DNS是否配置能访问https://ark.cn-beijing.volces.com

感谢已解决

Feb 22 '25 06:02 Week888

我先给个临时方案，直接改代码配置，能用model为bot开头的联网版。文件路径：app/constant.ts 在ByteDance配置里的ChatPath，增加bots

export const ByteDance = {
  ExampleEndpoint: "https://ark.cn-beijing.volces.com/api/",
  ChatPath: "api/v3/chat/completions",
};

修改为：

export const ByteDance = {
  ExampleEndpoint: "https://ark.cn-beijing.volces.com/api/",
  ChatPath: "api/v3/bots/chat/completions",
};

！副作用：该修改只能使用火山的应用，而不能直接使用模型

Feb 22 '25 10:02 mqfs

我先给个临时方案，直接改代码配置，能用model为bot开头的联网版。文件路径：app/constant.ts 在ByteDance配置里的ChatPath，增加bots
export const ByteDance = {
  ExampleEndpoint: "https://ark.cn-beijing.volces.com/api/",
  ChatPath: "api/v3/chat/completions",
};
修改为：
export const ByteDance = {
  ExampleEndpoint: "https://ark.cn-beijing.volces.com/api/",
  ChatPath: "api/v3/bots/chat/completions",
};
！副作用：该修改只能使用火山的应用，而不能直接使用模型

厉害大佬，已成功！其实谷歌的Gemini模型也自带联网搜索的功能，这是官方指南：“https://ai.google.dev/gemini-api/docs/grounding/?lang=python”，大佬有什么办法让gemini模型在nextchat上也能用上 grouding with google search嘛

Feb 24 '25 09:02 WadeQAQ

https 同问

Feb 28 '25 07:02 zengcrvip

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

https Same question

Feb 28 '25 07:02 Issues-translate-bot

https://ark.cn-beijing.volces.com/api/v3/bots/chat/completions# 末尾#号结束,#号不能缺失

Mar 14 '25 08:03 elmersky

我先给个临时方案，直接改代码配置，能用model为bot开头的以太网版本。文件路径：app/constant.ts 在字节跳动配置里的ChatPath，增加bots
export const ByteDance = {
  ExampleEndpoint: "https://ark.cn-beijing.volces.com/api/",
  ChatPath: "api/v3/chat/completions",
};
修改为：
export const ByteDance = {
  ExampleEndpoint: "https://ark.cn-beijing.volces.com/api/",
  ChatPath: "api/v3/bots/chat/completions",
};
后果：该修改只能使用火山的应用，而不能直接使用模型

使用成功，并且我自己打包了一个docker镜像 myiszz/nextchat-deepseek-r1-network:v1.0

Apr 30 '25 06:04 WishMelz