riwigefi / light-gpt

Light-GPT is an interactive website project based on the GPT-3.5-Turbo Model.

Hi there 👋

This is riwigefi, nice to meet you.

I am a front-end engineer. I love creating interesting things and trying out new technologies and frameworks.

light-gpt's Issues

Adding API Key to code and allow access via username password/ google login

thanks for making this app available! I like the functionalities and I would like to have it work without having to add my api key. do you think it is something possible?

Also, the mobile version sidebar is not displaying the full title, would it be worth hiding it and adding a button that makes it visible?

This key is associated with a deactivated account. If you feel this is an error, contact us through our help center at help.openai.com.

https://vercel.com/用这个搭建的，请求的ip是自己本地的还是vercel的。

刚刚一用就挂了

Feature query the balance corresponding to api key

$ curl -s  "https://api.openai.com/dashboard/billing/credit_grants" \
--header 'Content-type: application/json' \
--header 'Authorization: Bearer api_key' | jq   
{
  "object": "credit_summary",
  "total_granted": 18,
  "total_used": 0.178104,  # Amount used
  "total_available": 17.821896,  # Remaining Available Balance
  "grants": {
    "object": "list",
    "data": [
      {
        "object": "credit_grant",
        "id": "be32c280-ad4a-4e14-bc1a-077a32275f9e",
        "grant_amount": 18,
        "used_amount": 0.178104,
        "effective_at": 1677628800,
        "expires_at": 1688169600 # expiration time
      }
    ]
  }
}

Feature request: Better to have conversation saved

Thanks for your brilliant project! I really appreciate its simplicity and power.

In regards to start conversation from last context, would it be possible to add a feature that allows for saving conversations? This would be very helpful.

Dockerfile中的Node.js版本相关

1. 基于node:16.13.1打image的时候会报错，大概内容是16.13版本不支持pnpm，需要升级到更高版本
2. 本地升级nodeJs版本之后，基于master分支重新打了image，部署之后发现本人部署的与whynotisme/light-gpt UI方面出入较大，不太清楚whynotisme/light-gpt这个image是基于哪个分支打出来的
3. 建议增加个proxy功能，可以把https://api.openai.com写进配置文件，方面后续用反向代理域名进行替代

能不能让AI在聊天中根据它自己说的内容发图

比如AI描述了这样一个场景：
你决定进入通道，通道越走越窄，似乎是一条秘密通道。你小心翼翼地前行，不时听到远处传来的声音。终于，你来到了一扇巨大的石门前。门上刻着一些奇怪的符号，你不知道该怎么打开它。

然后自动代替用户发送（img-一扇巨大的石门，奇怪的符号）来生成图片

anyone can see you API_KEY in Network tab for dev tools

支持gpt4吗？而且docker的端口映射应该是3000:3000

自定义的角色持久化 + 支持保存多个角色，并提供切换

RT

I cannot change the model to Curie for some reason

when i change that to text-curie-001, it doesn't work. I am very new to this. Also, where do you store the api key? Thanks!

Image export function is not available

部署完域名解析到**还得开魔法..

部署完域名解析到**还得开魔法..