Basics

Build by hand

Build your first workflow without the AI: drag blocks, connect edges, configure steps in the Inspector, test each one, and activate.

The AI can build whole workflows from a sentence, but everything it does is ordinary canvas work you can do yourself. Building one small workflow by hand teaches you the grammar. We'll make: a webhook that receives an order and posts big ones to Slack.

1. Drag a trigger

Open Blocks in the left panel and drag Webhook (under START) onto the canvas. Every workflow begins with exactly one trigger-ish entry — the palette's START group.

Dragging the webhook block from the palette onto the canvas Blocks drag straight from the palette. The card shows the step's key config at a glance.

2. Add and connect steps

Drag an If (DECIDE) and a Slack app step (or Send email if you haven't connected Slack) to the right of it. Connect them: click-drag from the small circle on a node's output edge to the next node's input. If has two outputs — true and false — wire both; an unwired branch silently drops whatever flows into it.

Two connected steps with the true/false ports visible Ports are the small circles. Multi-output steps (if, switch, loop) label each port.

3. Configure in the Inspector

Click the If node. The Inspector opens on the right with the step's fields — for If, a single expression:

payload["body"]["amount"] > 1000

Two things to notice: the webhook's output nests the caller's JSON under body (see core concepts), and logic expressions use bracket syntax, not {{ }} templates.

4. Test before you trust

Open the Inspector's Test tab and hit Run now on the webhook — MadMax gives you the trigger URL and holds a sample window so you can send a real request (curl works fine). Each step you test pins its latest output, and downstream steps test against those pinned outputs — so you can prove the chain step by step with real data before anything goes live.

5. Activate

Hit Activate in the topbar. The canvas freezes into a snapshot, the webhook URL goes live, and the run appears in run history every time it fires. Keep editing freely — the live snapshot doesn't change until you activate again.

Where to go next