AI Agent

The AI Agent tool lets you run a full AI agent as a step in your workflow: it thinks through a task using a chat model, can recall past conversation history, and can call other actions as tools to get things done. This guide covers connecting an OpenAI account, adding the agent, and wiring up its chat model, memory, and tools.

Step 1: Get Your OpenAI API Key

Give the key a Name, choose the Project, set the Permissions, then click Create secret key.

Click Copy to copy your key — you’ll need to paste it into Zaplane next, since OpenAI won’t show it to you again.

Step 2: Create the Connection in Zaplane

In your WordPress dashboard, go to Zaplane > Connections, then click Create credential.

Search for and select AI from the list of apps

Set the Provider to OpenAI, paste your API Key, then click Save Connection.

Provider to OpenAI, paste your API Key and click Save Connection.

Step 3: Create a Workflow

Go to Zaplane > Workflows from your WordPress dashboard, then click Create Workflow to open the workflow editor.

Before adding an AI Agent tool, you need to set up a trigger first. Here is how to do it quickly:

Click Select an app under the Trigger box, then choose Apps.
Search for and select your trigger app — in this example, we are using aBlocks Contact Form.
Choose the Trigger Type that matches the event you want to watch for, then click Continue.
On the Configure tab, set up any required options, then click Continue.
On the Test tab, click Test Trigger to confirm it is working correctly, then click Save.

Once your trigger is saved, proceed to the next step to add the AI Agent tool.

Step 4: Add the AI Agent Tool

Click the + icon on your workflow canvas to open the Add Action panel, then select AI Agent from the list of built-in tools.

Step 5: Select the Action Type and Connection

On the Select tab, choose the Action Type, then choose the OpenAI connection you created under Select Connection. Click Continue.

Step 6: Configure the Agent

On the Configure tab, enter the Agent Instructions (the agent’s system-level behavior) and the Task / Message you want it to work on. Set the Max Tool Steps, choose the Response Format, and decide whether to Save turn to Memory. Click Continue when done.

Step 7: Test and Save the Action

On the Test tab, click Test Action to confirm the agent runs correctly, then click Save.

Step 8: Attach a Chat Model, Memory, and Tools

The AI Agent node shows three connectors underneath it — Chat Model, Memory, and Tools. Click each one to attach what the agent should use: a chat model action to generate its responses, a memory action to recall past history, and any actions you want the agent to be able to call as tools.

Step 9: Continue the Workflow After the Agent

Once the agent is fully set up, click the + after the AI Agent step to add your next action as normal, for example, creating a post in WordPress.

Step 10: Update and Activate the Workflow

Back on the workflow canvas, click Update to save your changes and keep the workflow Active.

Your Zaplane workflow now runs the AI agent automatically every time it fires — thinking through the task with its chat model, recalling memory, and calling its tools as needed.

Done!

The AI Agent tool is now live in your workflow and will run every time this step is reached.