Quick Start Guide¶
This guide will help you create your first NebulaFlow workflow in under 5 minutes.
Prerequisites¶
- VS Code 1.90.0 or later
- NebulaFlow extension installed
- (Optional) OpenAI API key and OpenRouter API key for LLM functionality
Step 1: Install the Extension¶
- Open VS Code
- Go to Extensions (Ctrl+Shift+X)
- Search for "NebulaFlow"
- Click Install
Step 2: Open the NebulaFlow Panel¶
- Open the Command Palette (Ctrl+Shift+P)
- Type "NebulaFlow: Open Workflow"
- Select "Open Workflow Editor"
Step 3: Clear the Default Workflow (Optional)¶
- The workflow editor opens with a default workflow (Git Diff → Generate Commit Message → Git Commit)
- To start fresh, click the Clear button (trash icon) in the sidebar actions bar
- Confirm the deletion to remove all nodes
- The canvas will be empty
Step 4: Add Nodes¶
- The left sidebar contains the Library with node categories (Agent, Text, Shell, Preview, Conditionals, Loops, Subflows)
- Drag an Agent node (LLM) onto the canvas
- Drag a Shell node (CLI) onto the canvas
Step 5: Connect Nodes¶
- Click the Agent node and drag to the Shell node
- Your workflow should look like:
Step 6: Configure Nodes¶
Agent Node (LLM)¶
- Click the Agent node
- In the property editor:
- Select a model (e.g., "gpt-4o")
- Enter a prompt: "Generate a simple greeting"
- (Optional) Adjust reasoning effort (minimal, low, medium, high)
- (Optional) Set a system prompt override
- You can also double-click the node body to edit the prompt directly
Shell Node (CLI)¶
- Click the Shell node
- In the property editor:
- Ensure Mode is set to "command"
- Enter the command:
echo Hello World! - (Optional) Configure shell, stdin, environment, safety, and approval settings
- You can also double-click the node body to edit the command directly
Step 7: Execute the Workflow¶
- Click the Run button (play icon) in the sidebar actions bar
- Watch the execution flow through the nodes
- View the output in the execution log
Step 8: View Results¶
- Check the execution panel for output
- Click on individual nodes to see their results
- The LLM node will show the generated greeting
- The CLI node will show "Hello World!"
Next Steps¶
Congratulations! You've created your first NebulaFlow workflow. Now explore:
- Workflows: Create more complex workflows with conditions and loops
- Nodes: Learn about different node types and their configurations
- Integrations: Connect to external APIs and services
- Advanced Patterns: Explore workflow design patterns for production use
Troubleshooting¶
Issue: Workflow doesn't execute - Check that all nodes are connected - Verify node configurations are complete - Look for error messages in the execution log
Issue: LLM node fails - Configure authentication for the selected pi provider - Check that the selected model is available - Verify the prompt is properly formatted
Issue: CLI node fails - Check that the command exists on your system - Verify command arguments are correct - Ensure proper permissions for the command