chatgpt-vercel: A TypeScript repository from byoungd

ChatGPT-Vercel

Introduction

Create a private ChatGPT website with one-click for free using Vercel, support muti text / images generation conversations. Powered by OpenAI API GPT-4/3.5 and Vercel.

Features

⚡ Deploy quickly and for free using Vercel
💬 Text conversation with the ability to switch models and set context length
🎨 Image generation conversation supports the DALL-E and Midjourney models. It also allows for the adjustment of image size and count.
🌈 Multiple preset prompts added to customize AI behavior
🌏 Switch between various languages, currently supporting Simplified Chinese and English
💭 Local chat history saved with search, import and export functionality

Live Demo

User Guide

You can access the online demo above or deploy it privately for experience.

Conversation

Click the icon at the top left to add a conversation, which has two types:

Text conversation:
- The model is switchable, supported models
- By default, it is a continuous conversation, and each sending will carry part of context.
- Supports adding preset prompts, type / or click the button at the bottom left to add.
- Supports model configuration, click the settings icon at the top right to configure.
Image generation conversation:
- The model is switchable, supports the OpenAI DALL·E model and Midjourney
- Does not support continuous conversation, and each sending will not carry the context.
- Directly input the image effect you want, for example: a cat.
- For model DALL-E, expend OpenAI tokens. The effective access time for the image link is 2 hours. Please make sure to save it in time if necessary.
- For model Midjourney, depending on the Discord configurations, image generation may take a while, with a default timeout of 5 minutes. Please be patient and wait.

History record

When Save all conversations is enabled in the global settings, it will be saved to local cache. By default, it will not be saved.

Operation

Press Enter to send
Press Shift+Enter to line break
Enter / to add preset prompts, and searching is also supported.

All settings

see Configurations

Getting Started

1. Create Project

Create Vercel project from a github forked project(recommand) or the following Deploy button.

2. Set OPENAI_API_KEY

There are three ways to set your OpenAI API Key:

Set Vercel Environment Variables OPENAI_API_KEY
Rename your .env.expample file to .env and set OPENAI_API_KEY
Set OPENAI_API_KEY within the page

Attention: For Vercel, all environment variables need to be redeployed to take effect.

3. Set Midjourney (optional)

If you want to use the AI drawing feature of Midjourney, you can configure the relevant Discord settings , including the following fields:

DISCORD_SERVER_ID
DISCORD_CHANNEL_ID
DISCORD_TOKEN

How to get ids and token:

You can visit midjourney-cookbook to get some samples about Midjourney prompts.

4. Keep code synchronized (optional)

see Sync Fork

Other deployment methods

Run pnpm build and pnpm run server. Refer: astro-node

Configurations

Deployment Configurations

All deployment configurations could be configured in the .env file or in Environment Variables of Vercel

Configuration	Default Value	Description
PASSWORD	-	Website access password
OPENAI_API_KEY	-	Key for API request, multiple keys are supported, separated by commas, how to generate
LANGUAGE	en	The default language of the website, including prompts. Supported languages: zh/en
API_KEY_STRATEGY	random	The scheduling strategy mode for multiple keys: polling/random
OPENAI_API_BASE_URL	api.openai.com	The default address of the requested api
DISCORD_SERVER_ID	-	Discord server id
DISCORD_CHANNEL_ID	-	Discord channel id
DISCORD_TOKEN	-	Discord token
DISCORD_IMAGE_PROXY	-	Discord image proxy url

Global Configurations

All global configurations will be stored locally

Configuration	Default Value	Description
OpenAI Api Key	-	Only a single key is supported. If it is configured on the page, the key in the environment variable will not be used
Language	en	The language of the website, including prompts. Supported languages: zh/en
Save all conversations	true	The conversation won't be lost after the page is refreshed
Temperature	1	The larger the value, the more random the answer, with a range of 0-2
Text Conversation Model	gpt-3.5-turbo	Model used in api request, supported models
Continuous conversations	true	Carry the context for the conversations
Number of historical messages carried	4	For continuous conversations, the number of historical messages carried
Image Generation Conversation Model	DALL-E	Supported models: DALL-E / Midjourney / Replicate
Number of generated images	1	The number of images generated in a single image generation conversation
Size of generated images	256x256	The size of a single image in image generation conversation
Discord Server Id	-	If it is configured on the page, the key in the environment variable will not be used
Discord Channel Id	-	Ditto
Discord Token	-	Ditto

Planned Features

Export functionality to export as markdown and images
Theme color switching support, currently defaulting to gradient purple
Audio conversation support
Image generation using other models

These are some of the planned features to be developed. Collaborations are welcome, and feel free to suggest other ideas by submitting issues.

Development

Requirements:

NodeJS v18 or higher
pnpm v7 or higher

Proxy

By default, a custom proxy is used to request openapi locally, and LOCAL_PROXY needs to be set in .env. There is currently no default proxy, so if you choose to set up a proxy, you assume the associated risks.
If there is no need for a proxy, you can disable local proxy by setting DISABLE_LOCAL_PROXY=true in the .env file. At this point, api.openai.com will be directly requested.

Development:

Run pnpm dev
Expose port, for example, when using in cloud ide, run pnpm start

Build:

Run pnpm build

Contribution

Any contributions are highly appreciated. Here are some tips:

To improve the translation or add a new language, modify the lang directory. If adding a new language, you will also need to modify src/utils/i18n.ts.
To improve or add new preset prompts, modify the prompts directory.
To optimize the API, modify the src/pages/api directory.
To optimize page interactions, modify the src/modules directory.
For new feature support, please open an issue directly.

byoungd/chatgpt-vercel