Numeric input validation failed: use the node's actual types and bounds

Symptoms and scope

Conversion failures and values outside min/max are related input-validation cases. Fix the field definition rather than silently clipping every value.

Error fragments for search; IDs, values and filenames may vary:

invalid_input_type
value_smaller_than_min
value_bigger_than_max

Source-supported context

Core validation performs type conversion and applicable min/max checks, while nodes may define additional rules. A form control that looks numeric does not prove the JSON value is valid. Source 1

Cases to distinguish

1. The API sends an empty string, unit-bearing text or locale-formatted number instead of the expected value.

2. A copied workflow relies on bounds from another node version.

3. Frontend assumptions differ from the backend definition or its custom validator.

Diagnostic sequence

The sequence below is an editorial procedure based on the cited context, not a diagnosis already confirmed for your environment.

Step 1. Record input_name, received_value and its original JSON type alongside the current node definition.

Step 2. Read this node's permitted range instead of inheriting steps, frames or dimensions from another model tutorial.

Step 3. Send the correct JSON type and an in-range value. Use actual true/false booleans rather than ambiguous strings.

Step 4. Apply bulk corrections only after confirming field meaning. Passing validation by truncating values indiscriminately may change the task.

Completion check

Types and ranges match the current definition, and execution behaves as intended after the change.

Limitations and cautions

Defaults are starting points, not universally optimal settings. Store version-specific limits rather than publishing a permanent numeric table.

Original sources

• ComfyUI execution.py validation — original record checked 2026-09-21.

Localization prepared: 2026-09-22. The source-check dates above were inherited from the existing article; this translation does not claim they were all rechecked today. No GPU run, minimum-memory measurement or runtime guarantee is supplied.

Related troubleshooting and guides

• POST /prompt returns 400: inspect node_errors before retrying
• Required input is missing: identify the exact node and field