网上有一些项目可以参考。比如最近的openai translator最近提交的支持web backend api 的MR

Do you plan on adding the features visible in the playground.
By this i mean things like the temperature, top_p, etc.

这样可以不挂代理直接反代

Sorry i'm just unclear what this is, what to do with it in ChatGPT, and how to run / install / make use of it in any way. Please update readme

可以像 https://github.com/lencx/ChatGPT 这个项目一样，使用网页接口吗？ 实在不想申请api key😂

I want to use the app with multiple accounts or API keys. Would the app be able to support this in the future? Thanks!

电脑是联想拯救者y7000

Hi,
thank you for your project, it looks amazing. I am trying to build and compile on linux (Arch).
I am new to Rust and AppImage technologies; I tried:
$> cargo build --release
$> cargo run
But, everything I get is: "Hello, NoFWL!"
Diving in the code, I have found just Javascript, no Rust...

Qestions:

1. is all the NoFWL code present in the repository?
2. how do I build and run it?

So I cloned the repo and attempted to build it using Cargo run build, it outputted a program that show's Hello World!, why does the repo don't build the binaries available for download?

I get a "Type Error" message every time I send a message.

windows下无法正常退出

同一样的一个问nofwl的回答似乎没那么智能，截图中对比了同一个问题在chatgpt和nofwl的回答

我选用的是GPT4 ，对话了大概10次左右，都正常，但是突然无法发送信息了，显示错误码400

Perhaps we can consider providing support for customizing the API URL and API URL Path. In some places, direct access to www.openai.com is not possible, so Cloudflare's worker function needs to be used for forwarding or May be a proxy may be needed.

当网络出现问题时，无法停止继续生成，也无法开始新的对话

I'd like to translate the app for the Italian language, but as it isn't open source I'm creating an issue to request access.

Since this is using our api keys to comunicate with the program, is there a possibility that we can modify which type of censorship we use on our end, i saw this when i looked at this.
[https://platform.openai.com/docs/api-reference/moderations/create]

It would be useful to see the answer stream in while it's being generated instead of waiting until the whole message is ready, especially when using GPT-4 which is slower.

很奇怪，获取到的结果只显示一部分，直接被截断了
不知道是nofwl的原因还是代理的问题？

Dock栏图标隐藏设置不生效

Hello! This is the best project I have seen that wraps OpenAI's API!
As the openai api can upload the whole chat history, how about adding the feature to allow users rewrite part of ChatGPT's responses and continue the chat?

not working anymore when using any input this error shows "Request failed with status code 429"

Hello brothers, is the response speed very slow? I will respond after more than 2 minutes of draw here...

My PC's IP cannot access https://api.openai.com, I wish the app to add an input field for users to input their own host, like a chatbox, on the settings page.

Just as the title said, I used to try "ChatGPT", but it didn't solve that I need to refresh it frequently. And I already get the API key, it would be good to use it instead of log in with openai account

今天第一次使用，出现了：timeout of 300000ms exceeded。使用了梯子和不用梯子都这个提示。

Is there a reason this needs elevated privileges/admin rights to install? Why can't it be installed via the local user or portable exe?

看到设置，版本可以选择3.5和4，我并非plus会员，然后在里面选择了4版本，不知道是不是真的能使用到4呢？

在linux下设置全局代理门槛较高且不便，期望可以添加一个设置socks5或者http代理的地方

我在ubuntu20.04系统上，出现如下提示：no-fwl: error while loading shared libraries: libssl.so.3: cannot open shared object file: No such file or directory

Hi

Add WebChatGPT: ChatGPT with internet access (chrome extension) like functionality in desktop version .

its check internet result and inject it in the prompt for more updated and accurate result.

regards

想请问一下，这个解压后怎么打不开？

显示为

开发者工具没有报错

另外打开的暗黑设置在已有对话的窗口显示稍有灰度增加，但新对话窗口为一片空白

我添加我的API Key，但是通过对话得知是GPT3，这是怎么回事呢

I installed it on Windows, and the prompts are saved on C:\Users\Username\.nofwl. I would rather save them on D:\nofwl, so if I format my computer unexpectedly, my data would be saved in a save location.

It would be fine to save the data alongside the application itself. I installed it on D:\nofwl. For example, the Sublime Text editor automatically saves its data on the current executable directory if it finds a data directory alongside the main binary. If you could follow the same strategy would be more than enough.

Hello community! I would like to suggest, if possible, adding the functionality of using the Bellard's Text Synth Server.
The server already has an API.

https://bellard.org/ts_server/

API doc:
https://textsynth.com/documentation.html

Example of usage:

curl https://api.textsynth.com/v1/engines/gptj_6B/completions
-H "Content-Type: application/json"
-H "Authorization: Bearer YOUR_API_KEY"
-d '{"prompt": "Once upon a time, there was", "max_tokens": 20 }'

Thanks for your time and help!

I always get this error, no matter my settings; however old chatgpt desktop app from you works fine

can't close app from tray win11

Is there a possibility that you add a black mode where the UI is even darker, kind of like what reddit has for mobile devices in "Dark OLED Mode"?

Hi, you can do the localisation using ChatGPT.
I tried it with the German version and the result was spot on.

点击新消息，然后再点击消息历史后，之前创建的聊天频道标题就会消失变成默认的你好

I can't access the config panel to change the context setting. When I first launched NoFWL, there was a config panel, but I can't access it anymore.

I love the app by the way, great work.

Request for Help: I am unable to bind my card and recharge for ChatGPT due to my location. If anyone is kind enough to contribute an API Key, please send me a private message. I don't have a large usage volume and mainly use it for testing when developing applications. If you feel that the API Key is causing any abnormal calls or want to revoke my usage, you can delete it at any time.

Alternatively, if you know of a stable recharge method, please let me know so I can recharge myself (I am in China and cannot obtain an overseas bank card).

Email: [email protected]

系统是win11，电脑型号是联想拯救者R2020 ，15.6寸笔记本。

I set it to use chatgpt 4, but after restarting the program, it comes back using chatgpt version 3.5.

On the development console, I see these requests:

Also, how can I know it uses version 4 instead of 3.5?