blrchen / chatgpt-lite

Fast ChatGPT UI with support for both OpenAI and Azure OpenAI. 快速的ChatGPT UI，支持OpenAI和Azure OpenAI。

ChatGPT Lite

English | 简体中文

Demo

Visit the ChatGPT Lite Demo Site

Features

ChatGPT Lite is a lightweight ChatGPT web interface developed using Next.js and the OpenAI Chat API. It's compatible with both OpenAI and Azure OpenAI accounts.

• Deploy a custom ChatGPT web interface that supports markdown, prompt storage, and multi-person chats.
• Create a private, web-based ChatGPT for use among friends without sharing your API key.
• Clear and expandable codebase, ideal as a starting point for your next AI Next.js project.

For a beginner-friendly version of the ChatGPT UI codebase, visit ChatGPT Minimal.

Prerequisites

You need an OpenAI or Azure OpenAI account.

Deployment

Refer to the Environment Variables section for necessary environment variables.

Deploy on Vercel

Click the button below to deploy on Vercel:

Deploy with Docker

For OpenAI account users:

docker run -d -p 3000:3000 \
   -e OPENAI_API_KEY="<REPLACE-ME>" \
   blrchen/chatgpt-lite

For Azure OpenAI account users:

docker run -d -p 3000:3000 \
   -e AZURE_OPENAI_API_BASE_URL="<REPLACE-ME>" \
   -e AZURE_OPENAI_API_KEY="<REPLACE-ME>" \
   -e AZURE_OPENAI_DEPLOYMENT="<REPLACE-ME>" \
   blrchen/chatgpt-lite

Development

Running Locally

1. Install NodeJS 20.
2. Clone the repository.
3. Install dependencies with npm install.
4. Copy .env.example to .env.local and update environment variables.
5. Start the application using npm run dev.
6. Visit http://localhost:3000 in your browser.

Running Locally with Docker

1. Clone the repository and navigate to the root directory.
2. Update the OPENAI_API_KEY environment variable in the docker-compose.yml file.
3. Build the application using docker-compose build ..
4. Start it by running docker-compose up -d.

Environment Variables

Required environment variables:

For OpenAI account:

Name	Description	Default Value
OPENAI_API_BASE_URL	Use if you plan to use a reverse proxy for api.openai.com.	https://api.openai.com
OPENAI_API_KEY	Secret key string obtained from the OpenAI API website.
OPENAI_MODEL	Model of GPT used	gpt-3.5-turbo

For Azure OpenAI account:

Name	Description
AZURE_OPENAI_API_BASE_URL	Endpoint (e.g., https://xxx.openai.azure.com).
AZURE_OPENAI_API_KEY	Key
AZURE_OPENAI_DEPLOYMENT	Model deployment name

Contribution

PRs of all sizes are welcome.

chatgpt-lite's Issues

copy & paste will add extra %20 to clipboard

sometime i ask gpt to give the code. directly copy & paste is ok, but some time format need readjust in the destination editor

click copy button at the output, and paste in the target editor, i see all [SPACE] are replaced with %20 as well as other ESCAPE code. for / \ and etc.

Hope we can fix that.

大佬你好，能否增加一个密码访问或者登陆访问控制，防止被其他人滥用

非常感谢大佬的源码，非常好用，响应速度真的很快，就是没有任何保护措施的话，地址泄露后可能会被滥用

目标目录中没有 app 子目录哈

Clone the repository and navigate to the app subdirectory in your terminal.

suggest set edge function in route.ts when deploy on Vercel

Problem

Deploying to Vercel under the Hobby plan, the duration limit is capped at 10 seconds, which can be problematic as OpenAI API responses often exceed this timeframe.

Solution

To circumvent the issue of receiving a 504 error, Introducing Edge Functions to route.ts by export const runtime = 'edge';, this helps avoid the timeout scenario.

Ref

• Vercel functions Max Duration: https://vercel.com/docs/functions/runtimes#max-duration

bug: Overwriting on Stream

When changing conversations while a response is streaming, newly selected conversation starts to receive the stream, possibly overwriting the original text.

To replicate:

Create two conversations.
In second convo, ask a long response like "Give me the plot of the Matrix movie in 5 paragpraphs". While response is streaming, switch conversations by selecting the other convo on the left hand side.

Model error message is not fully displayed

When a 400 error occurs on Azure OpenAI model calls, the detailed error message, such as "The response was filtered due to the prompt triggering Azure OpenAI’s content management policy. Please modify your prompt and retry. To learn more about our content filtering policies please read our documentation: https://go.microsoft.com/fwlink/?linkid=2198766" is not fully displayed in the error toast message box.

Ability to store past prompts and answers from ChatGPT

Is it possible to store past prompts and answers from ChatGPT?

Table does not render correctly in chat window

Please refer to the last section in the following screenshot. The table should render as an html table with border.

Messages not displaying on mobile

Tracking Messages not displaying on mobile issue reported by @ahlag

图片格式的文件上传能不能给加上呀

部署到vercel后如何设置为azure环境

我点击一键部署后，只配置了azure的环境变量，但是请求不到返回
请问还需要设置什么么？

Required OpenAI API on deployment Vercel

I try to set up with Azure OpenAI API but it's forcing me to use ChatGPT API. Could you make a change to that?

docker-compose只添加azure链接信息报错

您好，当我使用docker-compose部署时，配置文件仅添加了azure的相关参数。报错内容：[Error: The OpenAI API has encountered an error with a status code of 404 and message Not Found] 请问如何解决。

Integrate APIpie.ai

Founder here plugging my own product.

I tried to test our API with your app but found the UI doesn't permit changing the API Key / Base URL, it would be great to test but also APIpie.ai would be a great direct integration we have a huge range of models from different providers services and have also marked all Chat enhanced models (proper streaming and context memory) with subtype chatx so the models list can be filtered to only provide a list of models optimised for Chat experience, we are openai API compliant and we also update the models list frequently so would require no manual updating for new models if filtering on ChatX, you could further clean the list by removing duplicates if you don't mind what provider is serving the model.

https://apipie.ai/
https://discord.gg/fnVjanuCzB