Guides· Last checked October 28, 2026

ChatGPT says the conversation is too long. Here's how to keep going

When a ChatGPT chat gets too long, you may see this message: “You've reached the maximum length for this conversation, but you can keep talking by starting a new chat.”

You can't add to that chat anymore. The fix is to bring the state of your work into a new chat: the goal, what you decided, your rules, what is still open, and the next step.

Make a brief for a new chat

Paste the old chat and pick ChatGPT as the receiving AI. Free: 3 briefs a day per network. Your paste is not stored.

What the new chat needs from the old one

Goal

What you are trying to do, in one line.

Decisions and Current State

What you already chose and what is done.

Constraints and Preferences

Your rules: budget, tone, format, things to avoid.

Open Questions and Risks

What is still not decided.

Next Action

The one thing ChatGPT should do first.

Essential Context

Names, numbers, and facts ChatGPT can't guess.

What the message means

A single ChatGPT chat can only hold so much. Long pastes and long answers use up that room fast. When the chat is full, ChatGPT stops taking new messages there.

We could not find the exact size of that cap on OpenAI's help pages, and we found no setting that raises it.

Long chats often get worse before they stop. Old drafts and dropped ideas pile up, and ChatGPT can lose track of a rule you gave early on. That is why a fresh chat with a short brief can work better than the old chat did near the end.

Not the same as a message limit. If the message says you sent too many messages, that is a different limit. It is about how many messages you sent in a stretch of time, and it goes away after a while. This guide is about the “maximum length for this conversation” message.

Usually you can still open and read the old chat. If it won't open at all, OpenAI Support suggests asking for a copy of your data: Settings → Data controls → Export.

Fix 1: Branch from an earlier message

ChatGPT on the web can copy a chat up to any message into a new chat. This is called Branch in new chat. OpenAI Support suggests it as a workaround for chats that got too long.

Step 1.Scroll up.
Go back a few messages from the end of the full chat. If there was a big paste or a long side trip, pick a point before it.
Step 2.Branch.
Hover over that message, select More actions (⋯), then select Branch in new chat.
Step 3.Fill in what is missing.
The branch only has the chat up to that point. Tell it what happened after, in a few lines. Upload again any file you added after that point.

Branching buys time. It does not reset the chat for good. One person on OpenAI's forum branched the same long chat four times in a row. The first two branches each gave days of extra work. The next two filled up within hours.

Fix 2: Get a brief from the branch

A branch still has room to answer, so you can use it to write a brief of the work. Then move that brief into a clean new chat.

Step 1.Branch from an earlier message.
Do it the same way as in Fix 1.

Step 2.Ask the branch for a brief.

Send this prompt:

Write a handoff brief so a new chat can keep working on this without seeing our chat. Use these six headings:

1. Goal (one sentence)
2. Decisions and Current State (what is decided and done; if I changed my mind, give only the latest choice and say it replaces the old one)
3. Constraints and Preferences (rules, limits, tone, format, things to avoid)
4. Open Questions and Risks (what is not decided yet, or "None")
5. Next Action (one clear instruction)
6. Essential Context (only what is needed to continue)

Rules: Use only what we actually said. Don't guess or add anything. Keep names, numbers, and dates exact. List any files I need to upload again. No intro and no closing question.

Step 3.Add the end of the story.
The branch doesn't know what happened after the branch point. Add those last steps to the brief yourself.
Step 4.Check the brief.
Make sure every hard rule is there, like a budget, a word count, or “never do X.” Where you changed your mind, check that the latest choice is the one listed.

Step 5.Start a new chat.

Send this message, with the brief pasted below it:

I'm continuing work from an earlier chat. The brief below shows where it stands. Read it first. Then tell me in 2 or 3 sentences where we are and what you would do next. Wait for me before you start.

[paste the brief here]

Step 6.Upload your files again.
Files from the old chat don't come along with the text.

Fix 3: Write the brief yourself

If branching doesn't work for you, or you use the ChatGPT phone app, you can still write the brief from the old chat.

Step 1.Open the old chat.
Scroll to the part that matters most. Usually that is the last part.
Step 2.Write the six parts.
Use the list at the top of this page. Copy exact names, numbers, and rules from the chat. Keep it short. A brief that fits on one screen is enough for most work.
Step 3.Paste it into a new chat.
Use the starter message from Fix 2, Step 5.

Tip for long work: keep the chat inside a ChatGPT Project. Projects keep related chats, files, and instructions together. You can also save a reply, like your brief, to the project: open the reply's menu and select Add to project sources. Later chats in the project can then use it.

The faster way: KumiBrief

KumiBrief writes the brief for you from the text of the old chat.

Step 1.Copy the chat.
Open the old chat and copy it, or just the most recent part that matters.
Step 2.Paste it into KumiBrief.
Go to kumibrief.com, paste the chat, and pick ChatGPT as the receiving AI.
Step 3.Read, then copy.
KumiBrief writes the six parts. Read the brief, fix anything that is off, and paste it into a new ChatGPT chat.

It keeps your decisions and rules, and it uses your latest choice when you changed your mind.
It marks what is unknown instead of guessing.
It does not connect to ChatGPT. You copy and paste, so nothing needs a login.
The paste limit is 40,000 characters. A chat that hit the maximum length is likely longer than that, so paste the most recent part, plus any early message that set the rules.
The brief can be wrong. Read it before you use it.
Your paste is sent to Google's Gemini API to write the brief. KumiBrief does not store it. See the privacy page.

Make a brief for a new chat

Paste the old chat and pick ChatGPT as the receiving AI. Free: 3 briefs a day per network. Your paste is not stored.

How to avoid it next time

One task per chat. When the work moves to a new task, start a new chat with a short brief.
Hand off early. When a chat starts to feel slow or forgetful, ask for a brief while it still answers well. OpenAI's help pages suggest starting a new chat when a chat is long and ChatGPT gets slow.
Keep big files in a Project. Add them as project files instead of pasting them into one chat again and again.

Questions

Will paying for a higher plan make my chat longer?

Don't count on it. We found nothing from OpenAI that says a higher plan raises this cap. On Stack Exchange, one person upgraded to a paid plan after hitting the limit and said it had no effect.

Will a new chat remember the old one?

Not fully. With Memory turned on, ChatGPT can bring back some details from past chats. But OpenAI says Memory “does not retain every detail from every conversation.” For work in progress, paste a brief so nothing important is left to chance.

Why does ChatGPT get worse the longer I stay in one chat?

A long chat has many old versions of the same idea. ChatGPT can follow an old one or forget an early rule. A short, current brief in a fresh chat clears that up.

Why did my chat hit the limit so fast?

Everything in the chat takes up room, not just your messages. Long pastes and long answers fill it faster than many short questions.

Can I continue in Claude or Gemini instead?

Yes. The same six-part brief works in Claude, Gemini, and Grok. See [How to move a ChatGPT chat to Claude](/guides/move-chatgpt-to-claude). In KumiBrief, pick the AI that will receive the brief.

Sources

Menus in ChatGPT change often. We checked these pages on October 28, 2026.