Hi,
I am following the tutorial: OpenAI API request tutorial
the simplest way is to
client = OpenAI(api_key="paste-your-api-key-here")
now, you’ll have to go to platform.openai.com/api-keys to generate a key first.
also, do not share your code or upload your code with api_key="your-key-goes-here" as that will leak your key! if you plan to share your code: make sure to either take the key away or use something like the dotenv library.
something that helps is to copy and paste the documentation to chatgpt and ask it for help. Best of luck!
As @_thiago implied, it looks like you haven’t set your API key correctly. Also, best practice is to avoid hard-coding the API key (again as @_thiago says).
Thanks for your respond but I already did this and it give me even more error:
line 4, in
completion = client.chat.completions.create(
line 274, in wrapper
return func(*args, **kwargs)
line 742, in create
return self._post(
line 1270, in post
return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls)) and more errors.
Secondly, the reason I don’t use chatGPT because its code always give me this type of error:
You tried to access openai.ChatCompletion, but this is no longer supported in openai>=1.0.0
Hi @Idk ! Can you update your openai client? The latest version (right now) is 1.51.0, can you try pip install openai==1.51.0 and re-run it?
Hi!
This is my version before posting this error. I have tried many methods from the past forum post but nothing work for me.
Thank you
Hi @Idk . We are going down the rabbit hole here 
So your OP was an error because you didn’t supply a key in OpenAI() call, as @_thiago correctly stated.
The error You tried to access openai.ChatCompletion, but this is no longer supported in openai>=1.0.0 is related to incompatible use of chat completions API with that SDK.
The other error I am not sure about, it’s not actually giving me much info (possibly there is a more useful error message at the very end).
My suspicion is that it’s not a correctly setup Python environment using pyenv or virtualenv or whatever.
Hi @platypus. I think your suspicion may be valid because I don’t know how to correctly setup what I need.
And about the code:
\chatBot\main.py", line 4, in module
completion = client.chat.completions.create(
.venv\Lib\site-packages\openai_utils_utils.py", line 274, in wrapper
return func(*args, **kwargs)
.venv\Lib\site-packages\openai\resources\chat\completions.py, line 742, in create
return self._post(
.venv\Lib\site-packages\openai_base_client.py", line 1270, in post
return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls))
.venv\Lib\site-packages\openai_base_client.py", line 947, in request
return self._request(
.venv\Lib\site-packages\openai_base_client.py", line 1036, in _request
return self._retry_request(
.venv\Lib\site-packages\openai_base_client.py", line 1085, in _retry_request
return self._request(
.venv\Lib\site-packages\openai_base_client.py", line 1036, in _request
return self._retry_request(
.venv\Lib\site-packages\openai_base_client.py", line 1085, in _retry_request
return self._request(
.venv\Lib\site-packages\openai_base_client.py", line 1051, in _request
raise self._make_status_error_from_response(err.response) from None
Some questions that may help:
Edit: You answered some of that at the same time as I was writing the above, and I realised I made an error in my initial interpretation—my bad.
Those 1200 lines of code come from the openai package, not OP’s code. It doesn’t look like there is any wrapper here.
Looks like the response from the API side is not there, maybe something wrong with the API key? Or network connectivity?
Are you, by any chance, using a project-scoped key but not specifying the project ID?
Hi @caydennormanton
I am sorry I have no idea what does it mean. What I did was I go to this website as instructed: https://platform.openai.com/api-keys
Then I created new secret key → Owned by me → Default Project → Permission All. Then I copy the secret key and put in the code.
Use like: OpenAI(api_key=APIKEY, project=PROJECTID)
However, given that it’s the default project, I’m not sure this is necessary.
I used chatGPT free version for you, maybe it helps:
Absolutely! Here’s the complete setup for both Windows and macOS, including pip and the OpenAI package.
python --version.pip --version. If not, you can install it by downloading get-pip.py and running python get-pip.py.pip install openai in Command Prompt.print("Hello, World!"), and save it as hello.py.cd path\to\your\file and run python hello.py.python3 --version.pip3 --version. If not, you can install it by downloading get-pip.py and running python3 get-pip.py.pip3 install openai in Terminal.print("Hello, World!"), and save it as hello.py.cd path/to/your/file and run python3 hello.py.If you need further assistance with any of these steps, please feel free to use chatGPT!
I’ve been mulling this over, and the only thing I can think of is that it is either the API key configuration or network issues, and possibly for you to try using the same setup I have, which I could walk you through if you want, but will take me a few minutes to put together. I use python-dotenv, and the process involves setting up a .env file and using load_dotenv(find_dotenv()) to load the API key, project ID, and Org ID (you may not need all of these, depending on your account type).
If you want to give this a try, let me know and I’ll write up a quick guide.
Thank you for your help but I think I have an idea of what is happening.
There is one more error I did not show:
Spot on!
Let’s break it down:
It might be fixed—Try again, there was an issue.
@caydennormanton @Idk @_thiago this is awesome, true community effort here to solve a cryptic problem. Now we know!
