Error contacting the OpenAI API | Pawan.Krd | Page 1

cloud echo Oct 26, 2024, 11:08 AM

#

const response = await axios.post(
'https://api.pawan.krd/v1/chat/completions',
{
model: 'gpt-3.5-turbo', // Use 'gpt-4' or any other model you need
messages: [{ role: 'user', content: prompt }],
},
); doesnt work

cloud echo Oct 26, 2024, 11:53 AM

#

all i get is

📎 message.txt

#

this means end point is not working

tribal drum Oct 26, 2024, 12:16 PM

#

FIrst of, this API doesn't provide any OpenAI model.

#

Second, you need to set an API key in order to use the API, unless you are using cosmosrp.

cloud echo Oct 26, 2024, 12:19 PM

#

alrady using an api key provided by the bot

#

const axios = require('axios');
require('dotenv').config(); // Ensure you have this line if using dotenv

function createWindow() {
    const win = new BrowserWindow({
        width: 800,
        height: 600,
        webPreferences: {
            preload: `${__dirname}/preload.js`,
            contextIsolation: true,
            nodeIntegration: false,
        },
    });
    win.loadFile('index.html');
}

// Ensure you have your API key in your .env file
ipcMain.handle('send-message', async (event, message) => {
    try {
        const response = await axios.post('https://api.pawan.krd/v1/chat/completions', {
            inputs: message,  // Make sure this matches the expected payload
        }, {
            headers: {
                'Authorization': `Bearer ${process.env.API_KEY}`, // Use your API key
            },
        });
        return response.data.message; // Adjusted based on expected response structure
    } catch (error) {
        console.error('API Error:', error.response ? error.response.data : error.message);
        throw new Error('Error communicating with the API: ' + (error.response ? error.response.data : error.message));
    }
});

app.whenReady().then(createWindow);

// Handle window all closed event
app.on('window-all-closed', () => {
    if (process.platform !== 'darwin') {
        app.quit();
    }
});

app.on('activate', () => {
    if (BrowserWindow.getAllWindows().length === 0) {
        createWindow();
    }
});

#

as u can see nothing is wrong in here

tribal drum Oct 26, 2024, 12:20 PM

#

await axios.post(
            'https://api.pawan.krd/v1/chat/completions',
            {
                model: 'pai-001',
                messages: [{ role: 'user', content: prompt }],
            },
            {
                headers: {
                    'Authorization': 'Bearer Your API KEY', // Pur your API Key
                    'Content-Type': 'application/json'
                }
            }
        );

cloud echo Oct 26, 2024, 12:20 PM

#

as i send the mesage hi it jsut sits there and waits until no response from endpoint

#

is pai a model name?

#

we cant use gpt 3.5?

tribal drum Oct 26, 2024, 12:22 PM

#

No, you can't use the model "gpt-3.5-turbo" or any OpenAI model.

cloud echo Oct 26, 2024, 12:22 PM

#

just the pai?

#

anyothers with it?

tribal drum Oct 26, 2024, 12:22 PM

#

cloud echo just the pai?

https://discord.com/channels/1055397662976905229/1213171976407883866

cloud echo Oct 26, 2024, 12:24 PM

#

ok ill get back to u after testing these out

tribal drum Oct 26, 2024, 12:24 PM

#

Looks like that indeed the API is not returning any response, probably a down time.

cloud echo Oct 26, 2024, 12:24 PM

#

thanks for ur support

#

yeap 😛

tribal drum Oct 26, 2024, 12:24 PM

#

On line 22, the "inputs" propertie.

tribal drum Oct 26, 2024, 12:26 PM

#

tribal drum On line 22, the "inputs" propertie.

That's not how the body payload is supposed to be.

cloud echo Oct 26, 2024, 12:26 PM

#

i see

#

ill check that with the snippet u provided

tribal drum Oct 26, 2024, 12:27 PM

#

On line 28, response.data.message is also wrong.

#

The response payload is something like this:

{
  "id": number,
  "created": number,
  "model": string,
  "object": string,
  "choices": [
    {
      "finish_reason": string,
      "message": {
        "role": "assistant",
        "content": string
      }
    }
  ],
  "usage": {
    "prompt_tokens": number,
    "completion_tokens": number,
    "total_tokens": number
  }
}

cloud echo Oct 26, 2024, 12:31 PM

#

where can i get this payload in realtime with my api key?

tribal drum Oct 26, 2024, 12:32 PM

#

The response returned by the API should be like this.

#

As I said, looks like teh API is having a down time or something.

tribal drum Oct 26, 2024, 12:33 PM

#

tribal drum The response payload is something like this: ```json { "id": number, "create...

I think it won't return any response, but if it id, it would be like this.

cloud echo Oct 26, 2024, 12:34 PM

#

ok so i just had my first successfull answer to a hi

#

and the second too

#

okay now how many tokens pr day do i have?

tribal drum Oct 26, 2024, 12:39 PM

#

Every model spends 1 credit every X amount of tokens.

cloud echo Oct 26, 2024, 12:40 PM

#

that credit is taken by us? or?

#

API Response: {
id: 'chatcmpl-S5UGHftEBGdxsTQkiMBQqrjDyvZa',
created: 1729946348797,
model: 'pai-001',
object: 'chat-completion',
choices: [ { finish_reason: 'stop', index: 0, message: [Object] } ],
usage: { prompt_tokens: 5, completion_tokens: 155, total_tokens: 160 }
}

#

this is what i get..

tribal drum Oct 26, 2024, 12:40 PM

#

You have daily credits, your daily credits resets every day,

cloud echo Oct 26, 2024, 12:41 PM

#

and how many are those pr day?

tribal drum Oct 26, 2024, 12:41 PM

#

If you go to https://discord.com/channels/1055397662976905229/1064807234825113621 and use /info you'll be able to se it.

cloud echo Oct 26, 2024, 12:41 PM

#

thanks

tribal drum Oct 26, 2024, 12:41 PM

#

cloud echo API Response: { id: 'chatcmpl-S5UGHftEBGdxsTQkiMBQqrjDyvZa', created: 172994...

Now you can use dot notation to read the response, respose.choices[0].message.content

cloud echo Oct 26, 2024, 12:44 PM

#

so with one key i got from bot i can use any model?

#

or each key is assinged to each bot?

#

i mean i need different key for different bots

#

i means model

#

example 1 key = alll bots usage?

tribal drum Oct 26, 2024, 12:52 PM

#

You key has a tier, you can use all free models.

#

CosmosRP Pro, is not a free model, therefore you need to be a supporter to use it.

cloud echo Oct 26, 2024, 12:52 PM

#

when does the key expires like?

tribal drum Oct 26, 2024, 12:53 PM

#

However, you can use pai-001 and it's variants, and CosmosRP.

tribal drum Oct 26, 2024, 12:53 PM

#

cloud echo when does the key expires like?

They doesn't.

cloud echo Oct 26, 2024, 12:53 PM

#

so for example

#

const response = await axios.post('https://api.pawan.krd/v1/chat/completions', {
model: 'pai-001',

#

i can change the modal and url too? or the url remains the same for all bots

#

Developers: https://api.pawan.krd/pai-001/v1
Roleplay: https://api.pawan.krd/pai-001-rp/v1

tribal drum Oct 26, 2024, 12:56 PM

#

No, look.

#

AS you can see, this one has the model on the Endpoint.

#

So yes, you can change the URL, adding the model in the endpoint instead of in the request body in the model param.

#

But that won't change anything.

#

It's just the same.

#

And when using direct requests with axios, node fetch, cURL and etc, you need to add the /chat/completions to the endpoint.

cloud echo Oct 26, 2024, 12:59 PM

#

oki

#

i think the cosmosrp doesnt work https://api.pawan.krd/cosmosrp-it/v1/chat/completions

tribal drum Oct 26, 2024, 1:02 PM

#

Looks like CosmosRP is not working right now.

#

Everyone's having issues.

#

The other models should work fine.

cloud echo Oct 26, 2024, 1:03 PM

#

i see so the end point url needs to be changed like the one i gave just now? or willl it work with the normal one hs well?

#

https://api.pawan.krd/v1/chat/completions with this one?

tribal drum Oct 26, 2024, 1:04 PM

#

CosmosRP is the only model that doesn't work not using the model id in the endpoint.

tribal drum Oct 26, 2024, 1:04 PM

#

cloud echo https://api.pawan.krd/v1/chat/completions with this one?

Every other model, you can use this endpoint and just put the model id on the request body.

#

Note that if you put the model id in the endpoint, before the v1 it will always point to the same model.

cloud echo Oct 26, 2024, 1:05 PM

#

so can u give me a list that which model goes with which url? as if now em getting confused

#

cuz as u see if i use pai-001-light this with https://api.pawan.krd/pai-001-light/v1 it gives me Error communicating with the API.

#

thats why to understadn the model i need list

#

model and there url

tribal drum Oct 26, 2024, 1:08 PM

#

Sure.

   pai-001: https://api.pawan.krd/pai-001/v1/chat/completions
   pai-001-rp: https://api.pawan.krd/pai-001-rp/v1/chat/completions
   pai-001-light: https://api.pawan.krd/pai-001-light/v1/chat/completions
   pai-001-light-rp: https://api.pawan.krd/pai-001-light-rp/v1/chat/completions
   cosmosrp: https://api.pawan.krd/cosmosrp/v1/chat/completions
   cosmosrp-i: https://api.pawan.krd/cosmosrp-it/v1/chat/completions

tribal drum Oct 26, 2024, 1:08 PM

#

cloud echo cuz as u see if i use pai-001-light this with https://api.pawan.krd/pai-001-ligh...

As I said, you need to add the /chat/completions to the endpoint.

cloud echo Oct 26, 2024, 1:09 PM

#

hmmmm oooooooouuuuufffff a.I REALLY IS CONFUSING BUT learning then is enjoyable thansk for ur help today alot

#

really aprreactiated

tribal drum Oct 26, 2024, 1:09 PM

#

Yrw.

cloud echo Oct 26, 2024, 1:18 PM

#

though they still are not accurate as chat gpt its self right?

tribal drum Oct 26, 2024, 1:58 PM

#

No, they're not.

cloud echo Oct 26, 2024, 2:00 PM

#

are there models like chatgpt getting worked o riht now?

#

or currently these

tribal drum Oct 26, 2024, 2:24 PM

#

No.

cloud echo Oct 26, 2024, 11:28 PM

#

so successfully i have implemented it to a muktiplayer game and works perfect now i want to know that can we set custom commands with these modals? like to learn more about the game or tge server or multiplayer?

#

or just simppy make a custom command to do just the specific task?

tribal drum Oct 26, 2024, 11:30 PM

#

No idea what you meant by commands, however, in any way, that is something you would have to do by yourself based on what you want.

cloud echo Oct 26, 2024, 11:38 PM

#

can the modal remmember a prompt?

tribal drum Oct 26, 2024, 11:40 PM

#

As you can see in the example I gave you, messages is an array of objects.

#

Every object on this array is a different message.

#

If you include the "prompt" on this array it will "remember" it.

cloud echo Oct 26, 2024, 11:42 PM

#

example?

#

The response payload is something like this:

{
  "id": number,
  "created": number,
  "model": string,
  "object": string,
  "choices": [
    {
      "finish_reason": string,
      "message": {
        "role": "assistant",
        "content": string
      }
    }
  ],
  "usage": {
    "prompt_tokens": number,
    "completion_tokens": number,
    "total_tokens": number
  }
}
``` which part for ptompt

tribal drum Oct 26, 2024, 11:43 PM

#

Response.

tribal drum Oct 26, 2024, 11:44 PM

#

tribal drum ```js await axios.post( 'https://api.pawan.krd/v1/chat/completions',...

I'm talking about messages on the body payload.

tribal drum Oct 26, 2024, 11:45 PM

#

cloud echo example?

[
    {
        "role": "user",
        "content": "Hello, my name is Since."
    },
    {
        "role": "assistant",
        "content": "Hi, Since, nice to meet you!"
    },
    {
        "role": "user",
        "content": "Do you remember what is my name?"
    }
]

cloud echo Oct 26, 2024, 11:45 PM

#

await axios.post(
            'https://api.pawan.krd/v1/chat/completions',
            {
                model: 'pai-001',
                messages: [{ role: 'user', content: prompt }],
            },
            {
                headers: {
                    'Authorization': 'Bearer Your API KEY', // Pur your API Key
                    'Content-Type': 'application/json'
                }
            }
        );
``` but the content prompt gets the prompt only doesnt remmembers it

#

oh

tribal drum Oct 26, 2024, 11:45 PM

#

See what I meant?

cloud echo Oct 26, 2024, 11:45 PM

#

oki
got it

tribal drum Oct 26, 2024, 11:45 PM

#

That's how a chat history works.

#

Every message is sent as an object.

cloud echo Oct 26, 2024, 11:46 PM

#

so for how long do the modals remmembers it or if the session is over it will forget?

#

got it

tribal drum Oct 26, 2024, 11:46 PM

#

There's no "sessions".

#

It doesn't actually stores any data.

#

You're the one who sends every message.

tribal drum Oct 26, 2024, 11:47 PM

#

cloud echo so for how long do the modals remmembers it or if the session is over it will fo...

There's a limit of how much the model can process, that's all.

cloud echo Oct 26, 2024, 11:47 PM

#

oh..

#

i see ill need to work on it more tommorow

#

thanks again..!

#Error contacting the OpenAI API