WEBVTT

00:00.280 --> 00:08.400
Using it as a chat model to your agent, or making direct HTTP requests to Leipzig for specific tasks.

00:08.840 --> 00:11.000
So let's walk you through both methods.

00:11.360 --> 00:14.880
First is add the OpenAI chat model.

00:16.920 --> 00:19.800
So search for OpenAI.

00:22.040 --> 00:28.600
By default this points to OpenAI, but we will redirect it to Leipzig using the base URL.

00:29.880 --> 00:33.200
Next, go to open router and create an account.

00:33.680 --> 00:36.760
Once login, look for the DB six v3.

00:41.360 --> 00:43.080
And you will need the base URL.

00:43.720 --> 00:45.560
So click on API.

00:50.760 --> 00:52.640
And copy this base URL.

00:56.840 --> 00:57.400
Go back to.

01:02.760 --> 01:03.720
And paste it here.

01:06.240 --> 01:10.520
Next we need to replace the model name with Db6 specific identifier.

01:11.080 --> 01:13.400
So copy the model name from.

01:20.120 --> 01:22.680
Leipzig Leipzig.

01:23.120 --> 01:23.640
Chart.

01:29.920 --> 01:32.680
Switch to expression and paste it here.

01:35.000 --> 01:36.800
Now we have to set up credentials.

01:38.640 --> 01:40.320
So go back to open router.

01:43.600 --> 01:44.920
Create API key.

01:50.000 --> 01:50.800
Create key.

01:51.920 --> 01:52.760
Give it a name.

01:56.760 --> 01:57.400
Create.

02:01.040 --> 02:02.760
Copy your new key.

02:05.400 --> 02:09.200
Go back to N810 and paste it here.

02:12.480 --> 02:13.160
Now save.

02:17.640 --> 02:19.200
Don't worry about this message.

02:19.400 --> 02:20.760
It's just a bug.

02:20.800 --> 02:22.160
Everything will work fine.

02:22.960 --> 02:24.680
Now you can test the chat model.

02:25.560 --> 02:26.720
So click on the chat.

02:28.800 --> 02:30.000
You can ask it anything.

02:30.000 --> 02:30.440
Really.

02:32.640 --> 02:39.960
I ask it to write me a LinkedIn post about agents for small businesses in 2025.

02:41.360 --> 02:43.280
And as you can see, it's pretty good.

02:45.640 --> 02:49.480
Remember we did not provide any prompt to our agent.

02:49.960 --> 02:52.720
So we can achieve a lot better output.

02:54.000 --> 02:58.000
Now we will set up deep in HTTP request node.

02:58.240 --> 03:03.770
This method allows us to send prompts to seek and retrieve responses dynamically.

03:04.850 --> 03:07.290
So go to the deep sitcom.

03:14.250 --> 03:19.090
Create an account sign in and go to API platform.

03:24.450 --> 03:29.410
This will bring you to the API settings and here you can see your usage details.

03:29.690 --> 03:33.970
I just added two bugs for testing and it will go a long way.

03:34.290 --> 03:36.450
You will be able to process millions of tokens.

03:39.010 --> 03:45.090
You can top up your account using PayPal or debit or credit card.

03:46.690 --> 03:49.450
We need to create a new API key.

03:50.370 --> 03:52.770
So go to this section.

03:55.130 --> 03:56.450
Create new API key.

03:59.170 --> 04:01.570
Name it and copy it somewhere secure.

04:02.490 --> 04:04.650
You will need it for authentication.

04:11.770 --> 04:14.730
At HTTP request nodes to your canvas.

04:21.370 --> 04:22.690
Import curl.

04:24.410 --> 04:25.170
Paste it here.

04:26.410 --> 04:27.410
Click on the import.

04:29.970 --> 04:32.810
This will automatically set the method to post.

04:33.410 --> 04:37.970
Add the correct endpoint and pre-fill the required headers.

04:38.970 --> 04:40.170
So scroll down.

04:42.810 --> 04:46.450
And in the header section replace the placeholder.

04:49.650 --> 04:55.410
With your API key you generated earlier and make sure there are no extra spaces.

04:59.010 --> 05:00.090
It should look like that.
