Test a tool
Before you publish a tool, test it with real input. Open the tool from the registry and switch to its Test tab.
HTTP and Python tools can also be tested while you build them — see the live Send in HTTP tools and the test terminal in Python tools. This page covers testing a saved tool from its detail page.
Provide input
The Test tab shows a Test Tool card — “Execute this tool with sample input to verify it works correctly.”
Enter your input as JSON in the Input JSON box (it starts as {}, with a placeholder like {"param1": "value1"}). Supply a value for each input parameter your tool defines.
The input is a JSON object keyed by your parameter names — for a user_id parameter, enter { "user_id": 1 }.
Run it
Click Run Test. When it finishes, a result card shows:
- A Success or Failed status.
- The Execution time in milliseconds.
- The output as formatted JSON (or an error message on failure).
Test the unhappy paths too — missing required inputs, invalid values, and (for HTTP tools) error responses from the endpoint. Confirm the tool fails cleanly and returns a useful error.
Common issues
- “Invalid JSON input” — the Input JSON box doesn’t contain valid JSON. Check quotes, commas, and braces.
- Failed with a config error — a
{{config.xxx}}reference has no value. Add it under Configuration → Integrations (see Configuration & secrets). - Missing a required field — add every required parameter as a key in the JSON object.
- Slow or timing out — check the endpoint (HTTP) or the handler logic (Python); review the timeout you set on the Implementation step.