Mock data
Mock data is what a sandboxed node receives in place of a real response. Flero picks it from four sources, in a fixed priority order.
The four sources
| Priority | Source | What it is |
|---|---|---|
| 1 (highest) | Pinned | A payload you pinned on the node in the inspector. Applies in every mode, not only sandbox. |
| 2 | Snapshot | A curated response for a connector operation, maintained by the Flero team. |
| 3 | Execution history | The shape of a real response, found in a recent successful run in your workspace. |
| 4 (baseline) | Generated | A response built from the service's published API schema. |
When a node runs in sandbox, the first source that has something for that node and operation wins. If no source has anything, the node receives an empty object ({}).
Pinned
Pinned data is you saying "use exactly this for this node". It always wins, and the node does not run at all: no request is built and the mock service is never asked.
Pin from a run with Pin Data, or write your own with Edit JSON → Save. See Pinning data on a node.
Pinned data is the only source that keeps real values. Use it when you need a specific case: the unusual customer, the failed payment, the empty list.
Snapshot
Snapshots are curated responses that the Flero team ships with the product. They win over execution history and generated data for the same operation.
Execution history
When there is no snapshot for an operation, Flero looks through the 20 most recent successful runs in your workspace for a node of the same type and operation. The first clean output it finds is stored and reused for later sandbox runs.
Flero skips any candidate that looks like an error. An output that mentions words such as error, failed, invalid, unauthorized, timeout, not found or rate limit is passed over, even if the call succeeded. A clean output may be skipped by mistake. That is deliberate: it is better to fall back to generated data than to replay an error as if it were a success.
Generated
Generated data is the floor. It is built from each connector operation's published API schema, so every supported operation can return something correctly shaped. The values are generic.
Shapes, not values
Snapshot and execution-history data keep the shape of a real response, not its values. Every value is replaced with a placeholder before it is stored:
| Original value | Stored as |
|---|---|
| Any text | "string" |
| Any number | 0 |
| Any true/false | false |
| Object keys, lists, nulls | Kept as they are |
This keeps personal data and secrets out of mock data. It also means:
- An
Ifnode that tests a real value (amount > 100) may take a different branch in sandbox. - An expression that reads a field still resolves, because the field is there.
- If you need real values, pin the node.
What you see in the inspector
The node's Output tab shows whatever the mock data layer returned, exactly as for a live run. The inspector does not show which source won. If the output is placeholder-shaped ("string", 0, false), it came from a snapshot or execution history. If it is generic but varied, it was generated.
Tips and gotchas
- A stored execution-history response keeps being used. Once Flero finds a clean output for an operation, later sandbox runs reuse it.
- Set the operation on connector nodes. Mock data is matched by node type and operation. A node with no operation selected matches less precisely.
- Only objects count. A mock response must be a JSON object. A pin must be too.
- An empty object means nothing matched. Pin the node, or run the operation live once so a successful run exists for Flero to learn from.
Related
Found something out of date? This page lives in the Flero docs content set.