Documentation

Getting Started

Falcon Builder lets you create AI agents in three ways: use the AI Generator to generate workflows from natural language, start from pre-built templates, or build workflows manually with drag-and-drop nodes.

• Create an agent: Navigate to Dashboard → Agents → Create Agent
• Build a workflow: Click “Edit Workflow” to open the visual workflow canvas
• Add nodes: Drag nodes from the Node Library onto the canvas
• Connect nodes: Draw connections between nodes to define execution flow
• Configure nodes: Click any node to configure its settings in the right panel
• Test & deploy: Use “Test Workflow” to validate, then activate your agent

Core Concepts

Workflow Nodes Quick Reference

Built-in action and logic nodes available in every workflow.

Actions

• Log Message - Write messages to execution logs for debugging. Configure: log level (info/warn/error), message with variable interpolation.
• Delay - Pause workflow execution for a specified duration. Configure: duration value, unit (seconds/minutes/hours).
• Set Fields - Transform data by setting/mapping field values. Configure: field mappings (key → value), output variable name.
• Respond to Webhook - Send HTTP response back to webhook caller. Configure: status code, response body (JSON/text), custom headers.

Logic

• Condition - Branch workflow based on true/false evaluation. Configure: condition expression (JavaScript), true/false branch labels.
• Switch - Multiple conditional branches (like if/else if/else). Configure: multiple case expressions, default fallback branch.
• Loop - Iterate over arrays/collections. Configure: array variable, iteration logic, loop item variable name.
• Code - Execute custom JavaScript or Python code. Configure: language, code editor, output variable.

Variables & Expressions

Use variables to pass data between nodes. Reference variables using double curly braces: {{variable_name}}

Built-in Variables

• {{$json}} - Data from the previous node or trigger output
• {{$input}} - Webhook payload or trigger input
• {{$trigger}} - Original trigger data
• Custom output variables from previous nodes (e.g., {{$json.http_response}}, {{$json.ai_result}})

Expression Examples

• {{$json.user.email}} - Access nested object properties
• {{$json.data.items[0].name}} - Access array elements
• {{$json.email.from}} - Access email sender (Email Trigger)
• {{$json.email.subject}} - Access email subject (Email Trigger)
• Hello {{$json.name}}, your order is {{$json.status}} - String interpolation

Testing & Debugging

• Test Workflow - Click “Test Workflow” to execute with sample data before deploying
• Node Testing - Right-click any node → “Test Node” to test individual steps
• Execution Logs - View detailed logs for each execution in Dashboard → Executions
• Variable Inspector - See variable values at each step in execution logs
• Error Handling - Workflows stop on errors by default. Use try/catch patterns with conditions
• Replay Executions (Pro+) - Re-run failed executions with the same inputs

Best Practices

• Use Log nodes - Add Log nodes to track workflow progress and debug issues
• Name your nodes - Give descriptive labels to nodes for easier debugging
• Set output variables - Always configure output variable names for reusable data
• Test incrementally - Test each node as you build rather than testing the entire workflow at once
• Handle errors gracefully - Use Condition nodes to check for error states and branch accordingly
• Use credentials - Store sensitive data in Credentials instead of hardcoding in nodes
• Keep workflows focused - Break complex automations into multiple agents rather than one massive workflow
• Monitor executions - Regularly check the Executions page for failures or performance issues

AI Generator

Generate workflows from natural language descriptions. AI Generator is optional and uses credits from your plan.

How it works

• Describe what you want your agent to do in plain English
• AI Generator creates a complete workflow with nodes and connections
• Review and edit the generated workflow in the visual editor
• Refine by regenerating or manually adjusting nodes

Credits & Limits

• Free: 3 AI Generator credits/month
• Pro: 25 credits/month
• Business: 100 credits/month
• Enterprise: Unlimited

Bring Your Own LLM (BYO LLM)

Use your own LLM API key for unlimited AI generation without consuming platform credits:

• Step 1: Go to Dashboard → Credentials → Add Credential
• Step 2: Select credential type: OpenAI API, Anthropic API, Google API, or Ollama API
• Step 3: Enter your API key and give it a name
• Step 4: In the AI Generator modal, select your credential from the “LLM Provider” dropdown
• Result: Unlimited AI generation - tokens billed directly to your provider, no platform credits used

When using your own API key, the AI Generator bypasses credit limits entirely. You'll only pay for the tokens used from your LLM provider (typically $0.01-0.05 per generation).

Plan Features & Limits