Your First Workflow
This tutorial walks you through creating a simple image generation workflow. By the end, you'll understand the basics of adding nodes, connecting them, and running a workflow.
What We'll Build
A simple two-node workflow that generates an image from a text prompt:
[Prompt] → [Generate] → Generated ImageStep 1: Add a Prompt Node
- Look at the floating action bar at the bottom of the screen
- Click the Prompt button (or press
Shift + P) - A prompt node appears on the canvas
The prompt node has:
- A text input area for your prompt
- An expand button to open a larger editor
- A text output handle (small circle) on the right side
Step 2: Add a Generate Image Node
- Click the Generate button in the action bar (or press
Shift + G) - A Generate Image node appears on the canvas
- Drag it to the right of your prompt node
The Generate Image node has:
- Input handles on the left (for images and text)
- Output handle on the right (for the generated image)
- Provider and model selection dropdown
- Dynamic parameters based on selected model
- Custom configuration options
Step 3: Connect the Nodes
- Click and drag from the text output (right side of Prompt node)
- Drag to the text input (left side of Generate node)
- Release to create a connection
Connections only work between compatible types: text-to-text and image-to-image. You can't connect a text output to an image input.
Step 4: Write Your Prompt
- Click inside the prompt node's text area
- Type a description of the image you want to generate:
A golden retriever puppy sitting in a field of sunflowers,
soft afternoon lighting, professional photographyStep 5: Configure the Generator
- Provider: Start with "Gemini" (default)
- Model: Select a model like "nano-banana" for faster, cheaper generation
- Parameters: Configure model-specific options (aspect ratio, resolution, etc.)
Start with Gemini's "nano-banana" model while testing. You can explore other providers and models using the model search dialog for different capabilities and styles.
Setting Default Models
Save time by configuring default models for your project:
- Open Project Setup from the header menu
- Go to the Node Defaults tab
- Set your preferred default models for Generate Image and Generate Video nodes
- New nodes will automatically use these defaults
This is especially useful when you have a preferred model you use frequently.
Step 6: Run the Workflow
You have several options to run:
- Press
Cmd/Ctrl + Enterto run the entire workflow - Click the Run button in the header
- Click the play button on an individual node to run from that point
When running:
- Nodes execute in dependency order
- The current node shows a loading state
- Results appear when generation completes
Step 7: View Your Result
After generation completes:
- The generated image appears in the generate node
- Click the image to open a larger preview
- Use the carousel arrows to browse generation history
- Right-click to download the image
Adding an Output Node
For a cleaner workflow, add an Output node:
- Click Output in the action bar
- Connect the generate node's image output to the output node
- The output node gives you a dedicated preview area with download options
Saving Your Workflow
- Click the menu icon in the header
- Select Save (or press
Cmd/Ctrl + S) - Choose a location and filename
- Your workflow saves as a
.jsonfile
Node Banana auto-saves your work every 90 seconds when auto-save is enabled.
What's Next?
Now that you've created your first workflow, explore these topics:
- Core Concepts — Understand nodes, edges, and handle types
- Node Reference — Learn about all available node types
- AI Quickstart — Use templates and AI to build workflows faster