Know when an answer was cut short

You get complete answers you can trust, not replies that stop mid-sentence and look finished.

If an answer in chat hits its length limit, you will see a quiet notice that it stopped mid-answer. It will suggest you ask it to continue or narrow the question. Nothing failed, so it does not show as an error.

What to expect when an answer is cut short:

• In chat, the assistant tries once to rewrite a shorter answer. The rewrite starts on a new paragraph so it does not run on from the cut sentence. If that try is also too long, it reports honestly that it hit the length limit instead of pretending it finished.
• If a reply stops while it is still loading, the text you already have stays. You will see an error and a chance to retry, instead of a loading indicator that never ends.
• In a request's planning conversation, you will see a plain note that the reply stopped because of the length limit. It tells you what was kept and how to go on. For a blueprint or visual design step, it tells you nothing was saved and suggests asking for a shorter result.
• The notice stays when you reload or reopen the conversation, so a shortened reply is never left unexplained.

If you plan on WhatsApp, the next message you send after a cut continues that same attempt, so your change shapes the redo. You will only see a "ready" confirmation when a blueprint or design was actually saved. If an attempt ended with a question for you, your answer goes to that conversation instead of starting over.

Long planning threads also keep your latest instructions. The assistant always reads the newest messages and your original idea, and skips only the middle when it must. Learn more in Keep blueprints true to your latest decisions and Answer follow-up questions on requests.