Claude "reached its max length for a message": the fix
Claude paused because one reply hit its length limit, not your usage cap. Click Continue, or ask for the work in sections or as a file. Here's how.
"This response paused because Claude reached its max length for a message. Hit Continue to nudge Claude along." That notice means one single reply got too long, so Claude stopped writing partway through. It is not your usage limit and it does not lock you out. The fastest fix is to click Continue under the reply, or type "continue", and Claude picks up where it stopped.
If Continue fails, or you need something longer than one reply can hold, the fix is to change the shape of the request: ask for sections, or ask for a file. Both are below.
TL;DR: The pause is a per-reply output cap. Click Continue to resume. For long work, ask for one section per message, or ask Claude to put it in an artifact or a Word file instead of the chat.
The exact wording, and its variants
We could not open the claude.ai interface to capture the notice ourselves, so these are the versions people have reported, with where we saw each one:
- "This response paused because Claude reached its max length for a message. Hit Continue to nudge Claude along", with a Continue button. Quoted in GitHub issue #861, filed in April 2025 from the Claude web interface.
- "Claude hit the max length for a message and has paused its response. You can write Continue to keep the chat going." An older wording, quoted on the Claude Continue extension listing. This one asks you to type the word rather than click a button.
Both say the same thing. Anthropic's help center has no article for this notice, so treat the exact punctuation as reported, not official.
Why you see it
Claude works under two separate limits, and Anthropic's usage and length limits article draws the line clearly: usage limits "control how much you can use Claude across all your conversations", while length limits "control how long any single conversation can become."
This pause is a third, narrower thing: a cap on how much Claude writes in one reply. The developer side of Claude documents the same mechanism. Anthropic's stop reasons guide says a response can end because Claude "reached the max_tokens limit", and recommends you "continue the response".
What usually triggers it:
- asking for a whole report, chapter or long script in one go
- asking for a long piece of code, or a full rewrite of a long document
- asking for many items at once ("50 product descriptions")
How to fix it
Click Continue, or type continue
Claude resumes from where it stopped. On very long outputs you may need to do this more than once.
If it restarts instead, point to the last line
Type: Continue exactly from the sentence ending in [paste the last few words]. Do not repeat anything.
If Continue throws a length error, start a new chat
Paste the part you already have and ask Claude to write only what comes next.
That last step matters. In the GitHub report, clicking Continue in a long chat produced "Your message will exceed the length limit for this chat. Try shortening your message or starting a new conversation." At that point the chat itself is full, which is the conversation limit, and Anthropic's error messages page suggests starting a new conversation or breaking content into smaller chunks.
The Claude moves worth stealing, by email
Workflows tested on real jobs, sent only when they are worth it. Free.
How to get long outputs anyway
The reliable route is to stop asking for everything in one reply. Three ways, in the order we would try them:
1. Ask for sections. Start with: "First give me an outline with numbered sections. Then write section 1 only and stop. I'll say next for each section." Each reply stays short, and you can correct the direction after section 1 instead of after 5,000 words.
2. Ask for an artifact. Anthropic's artifacts article says Claude makes one for content that is "significant and self-contained, typically over 15 lines", and that you can "ask Claude to change the artifact" later. A long document then lives in one place beside the chat, and you can grow it section by section. Our Claude Artifacts guide shows how they work.
3. Ask for a file. Claude can create Word, Excel, PowerPoint and PDF files, per Anthropic's file creation article. Write "Create a Word document with the full report", and you download the result instead of scrolling a giant reply. It needs Code execution and file creation turned on in Settings > Capabilities, which is on by default for Free, Pro and Max.
A prompt you can paste this week
Take a long job you have been putting off, such as turning meeting notes into a full project brief. Attach the notes and paste:
I need a full project brief from these notes. Don't write it all at once.
1. Give me a numbered outline of sections first, then stop.
2. When I reply "next", write one section at a time.
3. When all sections are done, put the full brief in a Word document.
You end with a downloadable file, and no single reply comes near the cap.
How to stop it happening again
- Say the length you want. "Under 600 words" or "one page" keeps a reply well inside the cap.
- Split big batches. Ten descriptions per message, not fifty.
- Ask for changes, not reprints. "Only show the paragraphs you changed" beats "rewrite the whole document with these edits."
- Start fresh for big jobs. A long chat brings you closer to the conversation limit, which is what makes Continue fail. The usage limits guide explains how that differs from your 5-hour and weekly caps.
The honest limits
Anthropic does not publish how long a single reply in the Claude app can be, and it does not document this notice in its help center. So we cannot give you a word count, and anyone who does is guessing.
Continue is not perfect either. Claude can repeat a few lines, restart a section or lose the thread of a long table when it resumes. Check the seam where the two replies join before you copy the result anywhere. The reporter of issue #861 found the Continue-then-error loop frustrating enough to file a bug, and it was closed as "not planned", so do not expect a fix for that case soon.
Questions people ask
Does "reached its max length for a message" mean I hit my usage limit?
No. It means one reply got too long. Your usage limit is separate and covers all your chats over time. When you hit the usage cap, Claude stops answering until a reset, which is different from a reply pausing halfway with a Continue button.
How do I make Claude continue a response?
Click Continue under the paused reply, or type "continue" and send it. If Claude restarts instead of resuming, paste the last few words it wrote and ask it to continue exactly from there without repeating anything.
Why does Continue say my message will exceed the length limit for this chat?
The conversation itself is now full. That is the chat's length limit, not the reply cap. Start a new conversation, paste the part you already have, and ask Claude for only the next section. Anthropic's help center recommends a new conversation for this error.
How do I get Claude to write a really long document?
Ask for it in sections, one per message, or ask Claude to build it as an artifact or a Word file. That way no single reply has to hold the whole document. A numbered outline first makes each section easier to request.
If you want a short note when Anthropic changes limits like this one, the newsletter sends one when it matters.