Building and Testing Agents

Agents

Agents allow you to build the backend of your generative AI application. You can pick from a library of agents or create your own with your own models and data sets. The playground allows you to test the output of your agent against different prompts. The battleground allows you to test agents against each other to decide which one works best for your use case.

Before creating a agent, ensure that you have added at least one model and data source to your project. For more information, visit our section called “Model Lifecycle Management”.

Creating a New Agent

To create a new Agent:

1. Select the Project you want to create your Agent in.
2. Select the Agents tab.
3. Validate that you have added at least one Model and one Data Source to your Model Library and Data Sources.
4. Select “New Agent”.
5. Give the Agent a name in the top left.
6. Drag and drop nodes from the left side onto the canvas that you would like to use in your Agent.
7. Use the line drawer to connect the objects together.
8. Click “Save changes” or “Publish” to save your agent.

Test your agent in Agent Studio

You can now test your agent directly inside Agent Studio.
No more switching back and forth to a separate Playground while you’re iterating on your flow.

What you can do

• Inline Test mode – Chat with your agent without leaving Agent Studio.
• Playground Compare mode – Open the latest Draft and Published versions side by side to compare their behavior.

Both options live in the center-left controls on the Agent Studio canvas.

---

Inline Test mode (inside Agent Studio)

Inline Test mode lets you build, test, and refine in one place.

How it works

1. Open your agent in Agent Studio.
2. In the center-left canvas controls, click Playground.
3. A Playground panel opens on the left with a chat interface.
4. Type a message (for example, a realistic user input) and send it.
5. Watch how your agent runs through the steps on the canvas and returns a response.

The inline Playground uses the current configuration of your agent on the canvas, so every change you make is immediately testable.

When to use it

Use Inline Test mode when you want to:

• Quickly validate a single flow or prompt change.
• Tweak steps, retry, and iterate in a tight loop.
• Stay in a focused build environment without switching screens.

---

Playground Compare mode (Draft vs Published)

The original Playground Compare mode is still available when you need a more formal comparison.

What it does

Playground Compare mode:

• Opens your latest Draft and Published versions side by side.
• Lets you send the same input to both versions.
• Makes it easy to see how changes to prompts, tools, or flows impact responses.

How to open Compare mode

1. Open your agent in Agent Studio.
2. In the center-left canvas controls, click Compare in Playground.
3. The Playground opens with two panes:
   • Left: Latest Draft version
   • Right: Published version
4. Enter a message and compare how each version responds.

Editing a Agent

To edit a Agent:

1. Open the agent.
2. Make changes on the canvas.
3. Click “Save changes” or “Publish” to save your agent.

Debugging a Agent

To debug a Agent:

1. Test the Agent on the playground
2. Click the View Logs next to the output while testing, or navgiate to the Agent Executions feed on the left-side navigation and click on the Agent execution you are debugging.
3. Review the Logs for errors.

Using Your Agent

To use your agent in your application:

1. From the top navigation in Agent canvas, click the Settings icon.
2. Select "Interfaces".
3. Click "View API Info" next to the API interface.
4. The API endpoint for the agent will appear.
5. The headers of your API call should be:

• Content-Type: application/json
• X-API-Key: YOUR_API_KEY

6. View the API Documentation for the body of the request.

To Generate Your API Key

You can generate an API key from the agent's interfaces panel — click View API Keys next to the API interface — or from Settings → Developer → API Keys.

Either way you will name the key, set the project scope, and copy the key value before saving it. The full key is only shown once.

For complete interface setup guidance, see Managing Agent Interfaces.