Why a Dart model built from one JSON sample breaks on the next response

Generating a Dart model from a single sample API response is quick but often inaccurate, since one item cannot reveal every field's real shape. The example product shows how a model inferred from one item can wrongly treat an integer price as fixed, assume a discount is never null, and miss fields that only appear in other items. A later item with a fractional price or a null discount then triggers type-cast errors at runtime. Under sound null safety, fields present in every item should be required while those absent from some should be nullable, and numeric fields need careful handling. The author recommends building models from several real responses, including edge cases such as empty lists and missing values, and notes a free browser-based generator that merges sample items.