Skip to content
Docs
flero.ai

Sandbox mode, overview

Sandbox mode runs your workflow against mock data instead of real services. Toggle it on while iterating; turn it off when you're ready to do the real thing.

📸 Screenshot needed: sandbox__toggle.png, bottom execution chrome with the Sandbox toggle highlighted (switched on, with a yellow badge "Sandbox mode active" overlaying the canvas).

What it does

When sandbox mode is on, every external call from a Flero node is intercepted and answered with mock data:

  • HTTP nodes return canned responses instead of hitting the real URL.
  • Connector nodes (Slack, Salesforce, Stripe, …) return realistic mock objects.
  • Database nodes return rows from a mock store.
  • AI nodes return cached responses (no token spend).
  • Email and notification nodes log the message instead of sending.

Internal nodes (Transform, Code, If, Loop, etc.) run normally, they don't touch external systems anyway.

When to use it

  • Building a new workflow. Don't accidentally send 50 test emails to real customers.
  • Refactoring an existing workflow. Verify the structure before going live.
  • Demoing. Show how a workflow works without depending on a working network connection.
  • Replaying for diagnosis. Reproduce a failure path without re-triggering downstream side effects.

When not to use it

  • Final acceptance testing. At some point you need to hit the real thing.
  • Workflows whose logic depends on real-world data variance. Mocks are static; real APIs surprise you.
  • Performance testing. Mocked latency isn't real latency.

How to turn it on

Two places:

  • Per run, toggle Sandbox in the bottom execution chrome. Affects the next click of Run.
  • Default for this workflow, Workflow settings → Sandbox by default. Every run goes through sandbox unless explicitly overridden.

What changes during a sandbox run

Behaviour Production Sandbox
Webhook calls go out ✅ ❌ (mocked)
AI tokens are consumed ✅ ❌ (cached/synthetic)
Emails / SMS are sent ✅ ❌ (logged)
DB writes happen ✅ ❌ (recorded in mock store)
Workflow logic runs ✅ ✅
Expressions resolve ✅ ✅
Approvals pause ✅ ✅ (auto-approve toggle available)
Execution shows in history ✅ (regular) ✅ (tagged sandbox: true)

Sandbox executions in history

Sandbox runs appear in /logs like any other run, but are tagged sandbox: true. By default the monitoring dashboard excludes them; toggle Include sandbox to merge them in.

Found something out of date? This page lives in the Flero docs content set.