Back to Vibes

Core Concepts

Master the fundamental building blocks of Vibes including workflow components, step functions, variable management, context handling, and advanced automation patterns.

Understanding the fundamental components of Vibes is essential for creating effective, reliable workflows. Every Vibe consists of three essential identifying components that determine how it functions and when it gets triggered.

Name Component

The Name is the human-readable identifier for your Vibe that appears throughout the platform interface. This is what you and your team members will see in lists, logs, and management screens.

Good Examples

"Invoice Processing Workflow"
"Passport Expiration Alert"
"Customer Onboarding Process"

Poor Examples

"Workflow1"
"My Automation"
"Process"

Slug Component

The Slug is a unique, technical identifier used internally by the system to reference your Vibe. It must be unique across your entire organization and follows strict formatting rules.

Good Examples

passport-expire-alert
invoice-processing-v2
customer-onboarding

Poor Examples

PassportAlert
invoice_processing
workflow 1

Description Component

The Description is arguably the most critical component for Vibe functionality. This field enables the AI to determine when to trigger the Vibe during conversations through semantic matching.

Effective Description

"Takes an image attachment of a passport, extracts the expiration date, and sets up a scheduled email alert to be sent 1-month before expiration."

Poor Description

"Processes passports"

Workflow Steps and Functions

Vibes use a building-block approach where complex workflows are created by combining simple, pre-built functions in sequence. Each step performs a specific action and can pass its results to subsequent steps.

The Building-Block Philosophy

Rather than creating monolithic scripts, Vibes break down complex processes into discrete, reusable steps. This approach offers several advantages:

  • Modularity: Break complex workflows into smaller, focused components that can be maintained and updated independently
  • Debugging: Issues can be isolated to specific steps
  • Flexibility: Steps can be rearranged, added, or removed easily

How Steps Connect

Steps in a Vibe execute in sequence, with each step potentially using outputs from previous steps. This creates a data flow pipeline where information is processed, transformed, and passed along the workflow.

Basic Step Flow Example

Advanced Step Navigation

  • next_step_id: Jump to a specific step after completion
  • on_error_step_id: Jump to an error handling step if the current step fails
  • Conditional Logic: Use handleConditional to route based on data or conditions

Context and Variable Management

One of Vibes' most powerful features is its sophisticated context and variable management system. This enables workflows to maintain state, share data between steps, and build complex decision-making logic.

Understanding Context

Conversation Context

  • Complete conversation history between user and AI
  • User inputs and previous responses
  • Metadata about the conversation session

Workflow Context

  • Initial inputs that triggered the workflow
  • Results from all completed steps
  • Global variables set during execution
  • File attachments and user-provided data

Step Context

  • Specific inputs for the current step
  • Configuration parameters
  • Error states and retry information

Variable Types and Scopes

Context and Variable Examples

Advanced Variable Operations

Variable Operations Examples

Available Function Categories

Vibes comes with a comprehensive library of pre-built functions that handle common workflow tasks. These functions are organized into several categories:

AI and Processing Functions

aiExtractVariables

Extract specific information from text or images using AI

aiProcessing

Perform custom AI processing with flexible prompts

queryKnowledgebase

Search your organization's knowledge base

Communication Functions

sendResponse

Send messages back to the user via ChatBot

sendEmail

Send emails with custom content and formatting

promptUser

Request additional information from users

Data Management Functions

databaseExtraction

Query SQL databases with permission controls

setVariable

Create workflow-scoped variables

updateGlobalVariable

Create variables accessible across workflow steps

Control Flow Functions

handleConditional

Implement if/then logic and complex routing

loopFlow

Process arrays of data with sub-workflows

concludeWorkflow

Explicitly end workflow execution

Integration Functions

apiRequest

Make HTTP requests to external APIs

scheduleFlow

Set up time-based or recurring workflow execution

cancelScheduledFlow

Cancel previously scheduled workflows

Data Flow Patterns

Sequential Flow

Steps execute in order, each building on the previous

Step 1 → Step 2 → Step 3 → Complete

Conditional Flow

Workflow branches based on conditions and data

Condition → Path A or Path B → Continue

Loop Flow

The loopFlow function enables you to execute a defined sequence of workflow steps for each item in an array. This powerful function allows for bulk processing, automated repetitive tasks, and complex data manipulation workflows that need to operate on multiple records.

For Each Item → Process → Next Item

Parallel Processing

Handle multiple operations simultaneously

Multiple APIs → Combine Results

Error Handling Strategies

The apiRequest function provides comprehensive error handling for various scenarios:

Validation Steps

Include validation steps that verify data quality and completeness before proceeding with processing.

validation: required: true error_message: "Email is required!"

Conditional Error Detection

Use conditional logic to detect error conditions and route workflows to appropriate error handling steps.

on_error_step_id: "handle_api_failure"

Alternative Execution Paths

Create alternative paths that handle failure scenarios gracefully without losing work or leaving processes incomplete.

condition: then: "success_path" else: "fallback_path"

Human Escalation

Ensure complex situations receive appropriate human attention while maintaining workflow efficiency.

function: promptUser input: message: "Manual review required"

Error Handling Best Practices

  • Always validate inputs before processing to catch issues early
  • Design workflows to degrade gracefully when components fail
  • Provide clear, actionable error messages for users
  • Include fallback options for critical workflow paths
  • Log errors comprehensively for troubleshooting and analysis

Core Concepts Mastery

This variable injection system provides the flexibility to create sophisticated workflows that can process, analyze, and transform data dynamically while maintaining safety and performance standards.