Get Started
Key Concepts
A quick overview of the building blocks that power OAIZ automations.
Jobs
A Job is the basic unit of automation. It listens for an event, optionally applies trigger conditions, and executes an action.
- Input Event — What triggers the job (e.g., "GitHub issue created")
- Trigger Conditions — Optional filters (e.g., "only if label = urgent")
- Action — What the job does when triggered
- Output Event — Data published after execution for downstream jobs
- Attached Knowledge — Reference documents that provide context to AI agents
Actions
An Action is what a job does when triggered. OAIZ supports several action types:
- AI Agent - Simple — Single prompt/response with an AI model
- AI Agent - ReACT — Multi-step reasoning with tool use
- Deep Research — Extended research capability
- Webhook — Send data to external URLs
- Integration Actions — Actions in connected services (Slack, GitHub, etc.)
Chains
A Chain is a workflow that connects multiple jobs. Jobs in a chain pass data to each other through events.
- Visual builder on the Canvas
- Jobs connected by event edges
- Support for branching (fan-out) and merging (fan-in)
Events
Events are the data triggers that flow through OAIZ. They can come from:
- External integrations (GitHub, Slack, etc.)
- Previous jobs in a chain (output events)
Knowledge
Knowledge documents provide context to AI agents. Attach knowledge to jobs to give agents information they need.
- Rich text editor with version control
- Branching and merging like git
- PDF import support