1

Go to the Agent Composer

Navigate to the Agent Composer on the RELAI platform. This is where you’ll build and manage your AI agents.

2

Create a new agent

Click on the agent selection dropdown in the top-left corner and choose “Create Agent” to start building your agent. Give your agent a name to identify it later.

3

Add nodes to the canvas

You can add nodes to the canvas by clicking the ”+” button on the top-left corner of the canvas. For this quickstart, add the following nodes:

• Chat Trigger: This node will start the agent when a user sends a message in the chat panel (found on the bottom of the screen).
• Agent: This node will process the input from the Chat Trigger and generate a response using an LLM.
• LLM: This node will provide the language model configuration for the Agent node.
  You can rearrange the nodes on the canvas by dragging them.

4

Connect the nodes

Connect the nodes by dragging from an output port of one node to an input port of the next. Connect the Chat Trigger’s output to the Agent’s input, and connect the LLM node to the Agent’s LLM config port. After connecting, your canvas should look like this:

5

Configure the Agent node

Open the Agent node by double-clicking on it. In the parameters panel, set the parameters to desired values:

• Sytem Prompt: Enter a prompt that defines the agent’s behavior, e.g., “You are a helpful assistant.”
• Agent Input: Enter the input to the agent. To use the input from the Chat Trigger, set this to {input} (the default value). This works because the output of the Chat_Trigger node is passed as the input to the Agent node. You can also use the expression {@Chat_Trigger.output} to explicitly reference the output of the Chat_Trigger node. Read more about using Expressions in parameters.

6

Add LLM credentials

In the LLM node, you’ll need to add a credential that provides the API key for the language model you want to use (openai/gpt-5-mini by default). Start by clicking “Credentials” in the left sidebar. Then, click “Create new credential” > “Create Simple Credential”. In the modal that appears, give a name for your credential (e.g., openai-key) and select the credential type as LLM Credential. Set the provider to openai and enter your OpenAI API key in the “API Key” field. Finally, click “Create”. You should see your newly created credential in the table.Go back to your agent on the Agent Composer canvas, open the LLM node, and select the credential you just created from the “Select credentials” dropdown. The LLM node should now look like this:🎉 Congratulations! You’ve successfully set up a basic AI agent in the Agent Composer.

7

Chat with your agent

Now it’s time to chat with your agent! In the chat panel at the bottom of the screen, select the trigger as Chat_Trigger. Type a message into the input box and hit enter to send it to your agent. The Chat Trigger node will activate, passing your message to the Agent node, which will generate a response using the LLM you configured.You should see your agent’s response appear in the chat panel shortly after sending your message. Feel free to ask it anything or have a conversation!

​

Next Steps

• Learn more about Nodes and Edges in the Agent Composer.
• Understand how to set a node’s behavior using Parameters.
• The Agent Composer supports Expressions for dynamic data handling. This is a powerful feature and unlocks a great deal of flexibility.
• Explore more advanced conceptes like Execution flow.