"Prompt is too long" in Claude: what it means and the fix

In the Claude app, "prompt is too long" usually shows up as "Your message will exceed the length limit for this chat. Try attaching fewer or smaller files or starting a new conversation." It means your message, plus everything already in the chat, no longer fits in what Claude can read at once. The fastest fix: ask Claude for a summary, open a new chat and paste the summary in.

This is a length limit, not a usage limit. Waiting for a reset will not help, and paying for a bigger plan only helps if it gets you a model with a larger window. The steps below come from Anthropic's help center, checked October 22, 2026.

TL;DR: The chat is full. Get a summary, start a fresh conversation with it, and keep big files in a Project or send them in pieces.

The exact messages you might see

There are a few versions of this wall, depending on where you are:

• "Your message will exceed the length limit for this chat. Try attaching fewer or smaller files or starting a new conversation." This is the wording Anthropic documents in Troubleshoot Claude error messages, under length limit errors.
• "This conversation is too long to continue. Start a new chat, or remove some tools to free up space." Users report this one in chats, for example in this GitHub issue from May 2026. Anthropic's error article does not list it.
• "Prompt is too long" is the raw wording from the API and from Claude Code, often followed by a token count, as in this Claude Code bug report. People type it into Google because it is the short version of all three.

If instead a reply stops midway, that is a different limit, on how long one answer can be: see Claude's "max length for a message" fix. This guide covers the input side.

Why you see it: the context window, in plain words

Every chat has a fixed amount of text Claude can hold in view at once, called the context window. It is measured in tokens, small chunks of words. Everything counts against it: every message you sent, every answer Claude gave, every attached file, and the instructions for any tools or connectors you have switched on.

Anthropic's article on context window size on paid plans lists the current sizes per model. The newest models get 1M tokens in chat, several recent ones get 500K, and all other models get 200K, which Anthropic puts at about 500 pages of text or more. Part of the window is kept free for Claude's reply, so the usable space is a little smaller.

Per Anthropic's page on how usage and length limits work, paid plans with code execution enabled get automatic context management: as a chat nears the limit, Claude summarizes earlier messages and keeps going. That is why most people rarely hit this wall. Anthropic says rare cases, such as a very large first message, can still run into it.

So in practice you hit it for one of two reasons:

1. The chat grew too long. Weeks of back and forth, or many pasted documents, fill the window.
2. One message is too big. A huge PDF or several files at once overflow the space left, even in a fresh chat.

How to fix it

Fix 1: start a new chat with a summary

Anthropic's usage article says a new chat gives you a fresh context window. Here is the trick that keeps your progress (a common workaround, not an official feature):

If the full chat refuses even that short request, scroll up and copy the key parts yourself. Your old chat stays in your history either way.

Fix 2: move your files into a Project

A Project holds files as shared knowledge for every chat inside it, so you stop re-pasting the same documents. Anthropic's usage article names Projects as the way to work with larger amounts of information when you hit a length limit.

On paid plans, RAG for projects switches on automatically when a project's knowledge nears the context window. Claude then searches your files and pulls in only the relevant parts, which Anthropic says lets a project grow up to 10x. On the Free plan, project files still have to fit in the window. Our Claude Projects guide walks through the setup.

Fix 3: split the document

When one file is the problem, Anthropic's error article suggests splitting the content into smaller chunks and processing each separately, or summarizing key sections before sending. A workable pattern for a long report:

1. Send part one with: "Summarize this section in 10 bullets. I will send more parts."
2. Repeat for each part, in the same chat or new ones.
3. Paste all the bullet summaries into one final chat and ask for the overall answer.

Or ask Claude first which parts matter, as Anthropic also suggests, and send only those.

Fix 4: turn off tools you are not using

The second message above says "remove some tools to free up space", so tools take room too. Switching off connectors this task does not need frees space. Anthropic's error article does not cover this step; it comes from the message itself.

How to stop it happening again

• One task, one chat. Our prompts for work guide gives the same advice: start a new chat for a new task instead of piling onto one thread.
• Keep reference files in a Project instead of attaching them again in every chat.
• Trim before you paste. Send the chapter you need, not the whole book.
• Mind your usage too. Anthropic notes that long conversations use more of your allowance, so a fresh chat also stretches your usage limits.

What we could not confirm

We did not reproduce the error ourselves for this guide. Anthropic documents the "length limit for this chat" message, but not the "too long to continue" one, so we cannot say exactly when the app picks one over the other.

Claude.ai shows no token gauge, so you cannot see how close a chat is to the limit. Some users report the message in nearly empty chats (the GitHub issue above is one). If that happens, it is likely a bug, not your content: start a new chat or contact support.

And a summary is lossy. Claude can drop a figure or a caveat, so check the handoff note before relying on it.

Quick answers

Does upgrading fix "prompt is too long"?

Only if the new plan gets you a model with a larger window. The limit is per chat, not per day, so waiting does not reset it. Anthropic lists the paid-plan sizes per model, from 200K up to 1M tokens, on its context window page.

Will I lose my chat if I start a new one?

No. The old conversation stays in your chat history and you can reopen it to copy from. What you lose is Claude's working memory of it, which is why you paste a summary into the new chat.

Why does it happen on my first message?

The message, with its attachments, is bigger than the space left in a fresh window. Anthropic names very large first messages as the case automatic summarization cannot fix. Split the file, send fewer attachments, or put the files in a Project.

If you want short notes like this when Claude changes its limits, join the newsletter.