Core Concepts

Understanding the core concepts of ArcanFlows will help you build more effective AI-powered automations.

The Three Pillars

ArcanFlows is built around three interconnected components:

1. AI Agents

AI Agents are intelligent conversational interfaces powered by large language models (LLMs). They can:

• Understand Context: Process user messages and maintain conversation history
• Access Knowledge: Query documents and data you provide
• Take Actions: Execute tools and functions on behalf of users
• Delegate Tasks: Hand off work to specialized sub-agents

2. Workflows

Workflows are visual automation pipelines that orchestrate tasks. They consist of:

• Triggers: Events that start the workflow (webhooks, schedules, forms, etc.)
• Actions: Tasks to perform (send emails, call APIs, query databases)
• Logic Nodes: Control flow (conditions, loops, branches)
• Agent Nodes: Points where AI agents process data or make decisions

3. Smart Forms

Smart Forms are dynamic data collection interfaces that:

• Capture Input: Gather structured data from users
• Apply Logic: Show/hide fields based on conditions
• Validate Data: Ensure data quality before submission
• Trigger Actions: Start workflows or agent conversations on submit

How They Work Together

Loading diagram...

Example Flow:

1. A customer fills out a Smart Form to request a product demo
2. Form submission triggers a Workflow
3. The workflow invokes an AI Agent to qualify the lead
4. Based on the agent's assessment, the workflow routes to the appropriate sales rep
5. Notifications are sent via email, Slack, or SMS

Key Concepts

Tenants & Multi-tenancy

ArcanFlows is multi-tenant by design. Each organization (tenant) has:

• Isolated data and configurations
• Separate users and permissions
• Independent API keys and credentials
• Custom branding options (Enterprise)

Credentials Vault

Securely store and manage API keys, tokens, and other secrets:

• Encrypted at rest using AES-256
• Scoped access per agent or workflow
• Audit logging for all access events
• Automatic rotation support

Execution Context

When workflows and agents run, they have access to:

• Input Data: Variables passed from triggers or previous nodes
• Environment: Configuration and credentials
• Context: Metadata about the execution (timestamp, user, etc.)

Variables & Expressions

Reference dynamic data throughout your flows:

javascriptCopy
// Access form submission data
{{ input.email }}
{{ input.company_name }}

// Access previous node output
{{ nodes.http_request.response.data }}

// Use expressions
{{ input.amount * 1.1 }} // Calculate with tax
{{ input.name | uppercase }} // Transform text

Data Flow

Understanding how data moves through ArcanFlows:

Input → Processing → Output

1. Input: Data enters via triggers (form, webhook, schedule, event)
2. Processing: Workflows transform, enrich, and route the data
3. Output: Results are delivered via actions (notifications, API calls, storage)

Stateful vs Stateless

• Agents: Stateful within a conversation (maintain context)
• Workflows: Stateless by default (each execution is independent)
• Forms: Stateless (each submission is independent)

Security Model

Authentication

• Users: Email/password or SSO (Enterprise)
• API Access: API keys with scoped permissions
• Agents: Public (embed) or authenticated access

Authorization

Role-based access control (RBAC) with predefined roles:

Data Privacy

• All data encrypted in transit (TLS 1.3)
• Sensitive data encrypted at rest
• Configurable data retention policies
• GDPR compliance tools available

Next Steps

Now that you understand the concepts:

• Create Your First Agent
• Explore the Dashboard
• Build a Workflow