[Feature Request]: 关于二次开发如何接入新的API about chatgpt-next-web HOT 5 CLOSED

遵循 openai 的参数、响应结构和路由（v1/chat/completions），就可以套用了

from chatgpt-next-web.

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

Title: [Feature Request]: About how to access new API for secondary development

Problem Description

I would like to ask how to access the new API. An LLM is deployed on the server. I mainly want to add a UI interface to my LLM interface. Is there any specific operation or documentation?

Solution Description

none

Alternatives Considered

No response

Additional Context

No response

from chatgpt-next-web.

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

Follow openai's parameters, response structure and routing (v1/chat/completions) and you can apply it

from chatgpt-next-web.

你的LLM Api可以直接效仿 openai 的接口协议实现，然后，在部署nextchat应用的时候，在环境变量中设置BASE_URL指向你的LLM服务，在openai API协议的入参中，有一个model字段，标识的是你调用的模型名，这个model字段在你的LLM Api的接口实现中可以忽略掉

from chatgpt-next-web.

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

Your LLM Api can directly imitate the interface protocol implementation of openai. Then, when deploying the nextchat application, set the BASE_URL in the environment variable to point to your LLM service. In the input parameters of the openai API protocol, there is a model field, identified It is the name of the model you are calling. This model field can be ignored in your LLM Api interface implementation.

from chatgpt-next-web.