Skip to main content

How to Create Your First Workflow

Get started with AgentCanvas by creating your first workflow. This guide walks you through the process step by step, from choosing how to start to running your workflow.

Step 1: Choose How to Start

When you create a new application, you'll be asked how you want to start:

  • Start from scratch: Build your workflow from the ground up
  • Use a template: Start with a pre-built template from the marketplace
  • Let AI help: Describe what you want and AI will generate a workflow for you

For your first workflow, we recommend using a template or letting AI help you. Templates give you a working starting point, and AI can generate a complete workflow based on your description.

Step 2: Design Your Interface

Design how users will interact with your application:

  • Choose app type: Web app, API endpoint, or chatbot
  • Add components: Drag and drop UI components like inputs, buttons, and displays
  • Style your app: Customize colors, fonts, and layout
  • Preview: See how your interface looks as you build it

The interface builder lets you create the user-facing part of your application visually, no code required. You can always come back and modify it later.

Step 3: Build Your Workflow

Create the logic that powers your application:

  1. Add nodes: Drag building blocks from the node library onto the canvas
  2. Connect nodes: Connect nodes together to define data flow
  3. Configure nodes: Click on each node to set its configuration (API keys, prompts, etc.)
  4. Test nodes: Test individual nodes as you build to catch errors early
  5. Check costs: Use the cost estimator to see what your workflow will cost

Start with a trigger node (webhook, schedule, or manual trigger), add processing nodes (AI, data transformation, logic), and end with an output node (response, file, notification).

Step 4: Test and Run

Test your workflow before deploying it:

  • Test individual nodes: Run single nodes with sample data to verify they work
  • Test full workflow: Execute the entire workflow end-to-end
  • Check results: Review outputs from each node to ensure they're correct
  • Debug issues: Use execution logs to identify and fix problems
  • Optimize costs: Review cost estimates and optimize expensive operations

The workflow designer shows real-time cost estimates as you build, and you can test your workflow multiple times to ensure it works correctly before deploying.

Step 5: Deploy and Use

Once your workflow is ready, deploy it:

  • Save your workflow: Save your progress (auto-save happens automatically)
  • Publish: Make your application available for use (publishing is free)
  • Share URL: Share your application URL with users or integrate it into other systems
  • Monitor usage: Track usage, costs, and performance in your dashboard

Publishing your application makes it accessible via a URL or API endpoint. You can update it anytime, and changes are reflected immediately.

Quick Start Tips

  • Start with a simple template to learn the basics
  • Use AI to generate workflows—describe what you want and AI builds it
  • Test as you build to catch errors early
  • Check cost estimates before running expensive workflows
  • Save frequently—your work is auto-saved, but you can save manually too