How does asking for JSON differ from a schema-enforced response?
Answer out loud or in your head first, then compare.
Answer
Just asking “return JSON” is an instruction the model may follow wrongly. A JSON mode helps get a valid format, but it doesn't have to enforce our specific fields. A schema-conforming response constrains the structure as far as the given API supports.
It's the difference between “write something that can be read as JSON” and “return an object of a given shape”. I still have to handle refusals, interruptions and network errors and check the data's meaning. The schema also needs a way to mark missing information. If I force a price even though the document doesn't contain one, a valid structure won't stop the value being guessed.
Why it matters: Important when AI results are used automatically in code.
Possible follow-ups: Is every JSON Schema supported by every service?
How did it go? Be honest — “Not yet” puts the question on your review list.
Start with
- F-P01 · What is an LLM? — LLM fundamentals and AI engineering
- F-P03 · What are a token and a context window? — LLM fundamentals and AI engineering
Further reading
- Claude Platform: Structured outputs(opens in a new tab)Docs; accessed 22.10.2026. The schema contract, limits and validation; parameters depend on the API/SDK.