How-to

How to start a new ChatGPT chat without losing context

Your current thread is bloated and slow, but it holds decisions you can't afford to lose. Here is how to open a fresh chat and carry the important context with you — without dragging the whole mess along.

Starting over feels risky because you're afraid of losing the good stuff: the goal you finally nailed down, the decisions you argued through, the constraints you had to repeat five times. Good news — none of that has to be lost. You just need to carry the signal and leave the noise behind. That short summary is called a handoff.

What to carry over (and what to drop)

You do not need the whole transcript. You need four things. Anything outside these is usually noise that will just clog up the fresh chat.

The goal. One or two sentences on what you're actually trying to produce or decide. This is the thing the old chat probably drifted away from.
The locked decisions. Everything already settled — the approach, the tone, the format, the choices you don't want reopened. Write them as firm statements, not "we discussed."
The standing constraints. The rules you kept having to repeat: audience, length, what to avoid, must-haves. If you said it more than once, it belongs here.
The current state. What's done, what's in progress, and the single next step you want to take. This is where you pick the new chat back up.

The copy-paste handoff template

Fill this in, paste it as the very first message in a new ChatGPT chat, then ask your next question as if you'd never left.

You're picking up an ongoing piece of work. Here's the full context.

GOAL
- [What we're trying to produce or decide, in 1-2 sentences.]

DECISIONS ALREADY LOCKED (do not reopen these)
- [Decision 1]
- [Decision 2]
- [Decision 3]

CONSTRAINTS (always respect these)
- [Constraint 1 — e.g. audience, tone, length]
- [Constraint 2 — e.g. what to avoid]

CURRENT STATE
- Done: [what's finished]
- In progress: [what we're mid-way through]
- Next step: [the immediate thing to do now]

Confirm you've got this, then help me with the next step.

That last line matters. Asking it to confirm first gives you a quick sanity check that nothing important got garbled before you build on top of it.

Why a fresh chat is faster, not slower

A new chat starts with an empty budget, so everything you paste gets full attention. A short, clean handoff outperforms a long thread where your original goal has already scrolled out of view.

The two wrong ways people do this

Most people reach for one of these, and both cost you.

Wrong way #1: retype it from memory

You start a new chat and try to re-explain everything off the top of your head. The problem is you forget things — usually a constraint you'd repeated so often it felt obvious, or a decision you made hours ago. The new chat then cheerfully re-litigates a choice you already settled, and you're back to square one.

Wrong way #2: paste the entire old chat

The opposite instinct: copy the whole transcript into the new chat so "nothing is lost." But this recreates the exact problem you were escaping. You've refilled the budget with the same bloat, buried the goal under hundreds of lines, and the model has to wade through all of it again. A fresh chat stuffed with an old transcript is just a bloated chat with extra steps.

The handoff is the middle path: everything that matters, nothing that doesn't.

The faster way

Let Uncook write the handoff for you

Free & unlimited. No account.

Filling in that template means scrolling back through the whole thread to find every decision and constraint — the tedious part. Paste a share link to your ChatGPT or Claude conversation and Uncook reads the whole thing and writes the handoff for you: goal, locked decisions, repeated constraints and current state, already assembled. Skim it, paste it into a fresh chat, and keep going.

Uncook my chat →

Honest about your data: pasted text is analyzed in your browser; a share link is fetched once through our server to read the conversation, then discarded — never stored, never used for training. A share link makes the chat viewable by anyone with the URL; un-share it once you're done.