Zelaxy
Build AI-powered workflow automations on a visual, block-based canvas
Zelaxy is an AI-powered workflow automation platform built on a modular, block-based architecture. You connect blocks on a visual canvas — each performs a specific function — to build powerful automation pipelines without writing code.
Quickstart
Build and run your first AI agent workflow in five minutes.
Core Concepts
Blocks, connections, variables, and how data flows across the graph.
Guides
Complete example agent workflows — research, support, RAG, and more.
AI Models
Every provider and what each model can do.
Explore the reference
Blocks
274+ building blocks — AI agents, conditions, routers, loops, functions, and more.
Tools
249+ integrations — messaging, email, databases, search, AI/ML, and more.
Triggers
Webhooks, schedules, and service events that start workflows automatically.
Enterprise
SSO, self-hosted deployment, and the workspace security model.
How It Works
Every workflow in Zelaxy follows the same pattern:
- A trigger or starter initiates the workflow (user input, webhook, schedule, or event)
- Blocks process data — AI agents reason, conditions branch, tools fetch or send data
- Outputs flow between blocks using
{{blockName.outputField}}references - A response block (optional) returns the final result to the user
References always use double braces: {{blockName.field}} for another block's output, and {{ENV_VAR}} for secrets and API keys. Angle brackets like <block.field> do not resolve.
Core Concepts
| Concept | Description |
|---|---|
| Blocks | Self-contained processing units you drag onto the canvas. Each has inputs, configuration, and outputs. |
| Connections | Lines between blocks that define data flow. A block's output becomes the next block's input. |
| Variables | Reference any block's output with {{blockName.fieldName}} syntax — works everywhere. |
| Tools | External service integrations (Slack, Gmail, GitHub, etc.) that blocks can call. |
| Triggers | Events that start workflows automatically — webhooks, schedules, or service events. |
See Core Concepts for the full picture.
Example: Customer Support Bot
Here's a simple workflow that auto-responds to support emails:
- Gmail Trigger fires when a new email arrives
- Agent Block analyzes the email: classifies urgency, drafts a reply
- Condition Block checks
{{agent.content}}for urgency level - If urgent → Slack notifies the team; otherwise → Gmail sends the auto-reply
Want the full build? See the Customer Support Agent guide.
Explore the Platform
AI & Reasoning
The Agent block — multi-provider LLMs with tool calling, structured output, and memory.
Flow Control
Conditions, routers, switches, loops, and parallel execution.
Knowledge & RAG
Vector search across your documents for retrieval-augmented generation.
Build a Tool Call
Connect an Agent to real services like Slack, Gmail, and Notion.