Quickstart

Click on the “Add Agent” button in the top right corner of the Agents page.

Enter a name and an optional description of your Agent.

Select the Image Extraction Engine.

Each engine comes with two default settings for your Agents:

• Agent Input Definition: Defines the inputs that your Agent must accept when running a job. You can define the input key and its type as well as add or remove inputs. All defined inputs must be passed in a value when running the Agent. Learn more about Input Definitions.

Agent Inputs

• Engine Configuration: Configurations specific to your selected Engine. This sets things like which AI model or prompt instructions the Agent will use when running jobs. You can also dyanmically define the Engine configurations at runtime by using template strings of the Agent input keys. Learn more about configuring Engines.

Example of dynamic Engine configurations

In our example, let’s remove the model Agent input, fix gpt-4o as our model for this Agent, and leave the other Engine configurations as is.

Remove the model Agent input

Instead of setting the model dynamically at runtime using an Agent input, we fix it to always use gpt-4o in the configuration

Hit the Create button. Woohoo, you just built your first Agent! Now, let’s run it.

Paste in the following text for the instruction input field:

Describe the image in one sentence.
Output in the following structure:
```
{
  "description": "<description of the image>"
}
```

Then pick an image to upload to the image input field. You can use the one below.

san-francisco.jpg

Here is the filled-in Agent inputs:

Hit the Create button at the bottom. Hell yeah, your first Agent job has been kicked off!