How-to

How to hand off an AI conversation to a fresh chat

The real skill for working with AI over long projects isn't clever prompting. It's the handoff: moving your work from a tired chat into a fresh one without losing a thing. Do it well and you never fear the length limit again.

Why the handoff is the skill that matters

Every serious AI session eventually ends — it hits the length limit, or it rots (starts forgetting and repeating) well before that. When it does, most people either soldier on in a degrading chat or start over and lose their context. Both are avoidable. A good handoff lets a brand-new chat resume your work cold, as if it had been there the whole time.

A handoff is not a summary. A summary is for a human to read. A handoff is written for the AI to act on — a first message that makes the next chat continue your work without re-asking what you already settled.

The anatomy of a good handoff

Four parts. Miss one and the new chat starts guessing.

1. The goal

One line. What are you producing? Vague goals produce vague help. "A six-week go-to-market plan for a B2B scheduling tool" beats "help with marketing."

2. The locked decisions

The choices you already made and don't want reopened. This is what stops the new chat from relitigating settled ground. Example: "We're going with a single plan, not tiers. Don't propose alternatives."

3. The constraints you kept repeating

The single highest-value part, and the one people forget. If you found yourself reminding the AI of the same rule again and again — tone, length, format, a hard "never do X" — those are the first things a fresh chat won't know. Write them down verbatim.

4. The current state

Where the work actually stands. Paste the latest draft, the current version of the doc, or the exact question you were mid-answer on. Give the new chat a place to stand.

Frame it as a resume, not a briefing

Open the new chat with an explicit instruction to continue, not to start over: "Pick up exactly where we left off. Don't re-ask anything below — it's settled." Without that line, models tend to politely re-open everything you just told them.

A template you can steal

Pick up exactly where we left off. Don't re-ask anything below — it's settled.

GOAL: [one line — what we're producing]

LOCKED DECISIONS (don't reopen):
- [decision]
- [decision]

CONSTRAINTS (always follow):
- [rule you kept repeating]
- [rule you kept repeating]

CURRENT STATE:
[paste the latest draft or where you stopped]

MY ROLE / YOURS: [e.g. I decide scope and copy; you draft and pressure-test.]

Continue from here. Be direct, no recap.

Common mistakes

Pasting the whole old chat. This just recreates the bloat that killed the first one. Distill, don't dump.
Leaving out the repeated constraints. They feel obvious to you now, so you skip them — and the new chat immediately breaks them.
Writing a summary instead of instructions. "We discussed pricing" tells the AI nothing. "Pricing is locked, don't reopen it" does.
Forgetting the "don't re-ask" framing. Without it, you'll spend the first few messages re-confirming things you already handed over.

The honest shortcut

Once you've done a few handoffs by hand, you'll notice it's the same job every time — and the same part (those repeated constraints) is the one you keep under-doing. That repeatable, error-prone job is exactly what Uncook automates.

Skip the manual work

Get a handoff written for you

Free & unlimited. No account.

Paste a share link to your ChatGPT or Claude conversation. Uncook reads the whole thing and assembles the handoff — goal, locked decisions, the constraints you kept repeating, and the current state. You skim it, tweak anything, and paste it into a fresh chat.

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.