I wrote a small c# program to access text-davinci-002 and when I ran it I got a 404 error in response. I then tried a more direct approach and used POSTMAN to submit my GET request:
https://api.openai.com/v1/text/generate?prompt=What is the best tempurature for baking short bread cookies?&model=text-davinci-003&temperature=0.5&max_tokens=1024
and that too returned a 404 Not Returned error.
I verified the outbound using Fiddler4 and I even asked chatgpt if my url was correct (it said yes) and I do have a valid APIKey.
When I did a ping of the url I got timeouts?
and this whether on or off vpn.
My question is, is that the right base url (api.openai.com) or should I use something else? If it was authentication or a bad syntax I figure I get a different error. The c# code for the request I got from chatGPT so maybe it is outdated?
As you can see from the other post, ChatGPT isn’t the best way to learn coding for a specific API.
It’s best to stick with the docs and API examples.
[
## Create completion
post https://api.openai.com/v1/completions
Creates a completion for the provided prompt and parameters
3 Likes
Well, that was embarrassing, but then not that far off from some of the answers I may read on stackoverflow at times. perhaps what is slightly frustrating are that the examples are only for node.js and python. I should have c# web requests coding memorized by now, but as a data integrator just coming off SOAP, it is not build in yet. Let’s see if I can get this working now.
(Also, I normally don’t like to ask it “Please write me code I should understand”, but I was trying to save time).
1 Like
No worries. Yours wasn’t half as bad as some of the requests we get around here! 
In all seriousness, feel free to post something if you run into trouble. We’re all still learning …
1 Like
I still think, man that seemed so right (lol). Okay, using POSTMAN I get that the body is JSON established parameters. I have the body set to application/json and I am sending this:
{
“model”: “text-davinci-002”,
“prompt”: “What is the best tempurature for baking short bread cookies”,
“max_tokens”: 100,
“temperature”: .5,
“n”: 1,
“stream”: false,
“logprobs”: null,
“stop”: “\n”
}
My response is this:
{
“error”: {
“message”: “you must provide a model parameter”,
“type”: “invalid_request_error”,
“param”: null,
“code”: null
}
}
with my actual request (GET) looking like this:
GET https://api.openai.com/v1/completions HTTP/1.1
Authorization: Bearer
Content-Type: application/json
User-Agent: PostmanRuntime/7.29.2
Accept: /
Cache-Control: no-cache
Postman-Token: 9ebb4e96-5438-4e28-9d42-eb023993ad50
Host: api.openai.com
Accept-Encoding: gzip, deflate, br
Connection: keep-alive
Content-Length: 226
{
“model”: “text-davinci-002”,
“prompt”: “What is the best tempurature for baking short bread cookies”,
“max_tokens”: 100,
“temperature”: .5,
“n”: 1,
“stream”: false,
“logprobs”: null,
“stop”: “\n”
}
I can get a list of models, 002 is in the list. If I can get this working in POSTMAN translating it to c# is no issue. I fell like I am missing something from the documentation.
Needs to be a POST request not GET, I believe…
[
Create completion
](OpenAI API)
POST https://api.openai.com/v1/completions
Creates a completion for the provided prompt and parameters
(Seems I cannot read after 40+ years programming. Actually I’m just frustrated and was rushing)
I did get a response, but
{
“error”: {
“message”: “You exceeded your current quota, please check your plan and billing details.”,
“type”: “insufficient_quota”,
“param”: null,
“code”: null
}
}
I guess I need to read up on what that means now though is seems like am Amazon throttling thing. Thank you for your guidance.
There is a quota. You might need to fill in your billing details too…
Looks like you’re getting closer, tho!
Slow and steady…I was just going to go there and I am set up as Personal. According to them it’s 20 a min and I did not have one yet.
Yes. Users on the Free trial have an organization-level rate limit set to 20 requests/minute and 150,000 tokens/minute.
I tried added Personal as the OpenAI-Organization but it did not seem to like that. Perhaps I’ll try a new API Key (sigh)
