Mesh Canvas

The Mesh Canvas is the core of the flow editor. It represents each step in your flow as a node, connected by edges that define execution order. Every change on the canvas is reflected immediately in the YAML, and vice versa.

Open the canvas by navigating to any flow and clicking Edit, or creating a new flow from the Flows page.

Canvas navigation

A minimap in the corner shows the full flow at a glance. Use it to navigate large flows quickly.

Node types

The canvas uses four node types:

Action nodes represent individual test steps. Each node shows the step ID, action type, and current status during execution. The top edge of each node is colour-coded by protocol category:

Section headers visually separate the three execution phases: Setup, Steps, and Teardown. Setup and teardown steps run before and after the main steps regardless of pass/fail.

Condition nodes represent condition (if/else) steps. They show two outgoing edges: one for the then branch and one for the else branch.

For-each nodes represent for_each loops and show nested steps inside a container.

Adding steps

Steps are added from the action palette on the left side of the editor. Actions are grouped by category. Click any action type to add a new node to the canvas, then drag it into position.

You can also drag directly from the palette onto the canvas.

Connecting steps

Draw an edge from the bottom handle of one node to the top handle of the next to define execution order. The flow runs top-to-bottom following the connections you draw.

For condition nodes, connect the then and else outputs to the appropriate branches.

Properties panel

Select any node to open the Properties panel on the right. It has six tabs:

General

• Step ID (must be unique within the flow)
• Display name and description
• Timeout override (e.g. 30s)
• Run condition — an expression that gates whether this step executes at all (e.g. ${env.STAGE} == "staging")
• Retry configuration — max attempts, delay between retries, backoff strategy

Config Action-specific configuration form. Fields change based on the action type:

• HTTP: method, URL, headers, auth, request body, timeout
• Database: connection string, SQL query, parameters
• Kafka: brokers, topic, key, value, consumer group, timeout
• Redis: host, port, key, value, TTL
• And so on for all 54+ supported action types

Assert Define assertions using expr-lang expressions. All assertions must pass for the step to succeed. Examples:

status == 201
body.id != ""
messages[0].value.order_id == "{{order_id}}"
rows[0].status == "confirmed"

Output Map response values to variables using JSONPath. Variables are available to all subsequent steps via {{variable_name}} syntax:

user_id  →  $.body.id
token    →  $.body.access_token

Error Configure error handling: retry policies, fallback steps, and whether a failure should stop the entire flow.

Comments Leave notes on individual steps for collaborators. Comments are persisted and visible to all workspace members with access.

YAML mode

Toggle between visual and YAML mode using the tabs at the top of the editor. The YAML editor shows the full flow definition in the same format used by the CLI.

Edits in YAML mode are parsed and reflected on the canvas when you switch back to visual mode. Syntax errors are shown inline before the switch is allowed.

The YAML format is identical to what you'd write by hand — no lock-in to the visual editor.

Auto-layout

Use the Auto-layout button in the toolbar to automatically arrange nodes into a clean top-to-bottom flow. Useful after adding many steps manually or importing a YAML file.

Setup and teardown

Every flow has three sections available on the canvas:

• Setup — steps that run before the main flow, regardless of outcome (e.g. create a test user, seed data)
• Steps — the main test scenario
• Teardown — steps that always run after the flow completes, even on failure (e.g. delete the test user, clean up state)

Section headers on the canvas visually separate these three areas.