Carry a project into a new chat, model, or teammate without the context evaporating — capture the state, distill what's worth keeping, and rebuild it as durable context on the other side.
Context is the expensive part of working with AI, and it's the first thing you lose. A chat hits its limit, you switch models, you hand the project to a colleague — and the decisions, constraints, and half-finished reasoning vanish, living only in a thread nobody can fully reread. Starting over is the tax everyone pays and nobody budgets for. A deliberate handoff captures what the work is, what's been decided, and what comes next, in a form the next session can absorb in one read.
Each step uses an existing NewPrompt tool, pre-filled by a matching resource. Open the resource to read it, or jump straight into the tool with the inputs ready.
Capture the current state
Pull the live state out of the conversation — what's done, what's decided, what's still open — into a structured handoff instead of a wall of chat history.
Distill what's worth carrying
Long threads bury the few instructions that actually matter. Extract the working prompt, the conventions, and the preferences the next session should keep, and drop the rest.
Re-establish it as durable context
Turn the handoff into standing project context the new session reads first — so the model starts informed instead of asking you to re-explain the project.
Review the handoff for completeness
Check the handoff package for missing decisions, assumptions, constraints, files, risks, and next actions before another chat or teammate continues the work — so gaps surface now instead of after the next session is already stuck.
The new session — a fresh chat, a different model, or a teammate — opens already knowing the project's state, decisions, and conventions, with the handoff checked for gaps before anyone relies on it. The work continues instead of restarting.
Use handoff to preserve context you already built so it survives a jump to a new session, model, or teammate; use onboarding to learn an unfamiliar codebase from scratch. Handoff's four steps capture, distill, and re-establish standing context; onboarding builds understanding you don't yet have.
No. The Context Handoff Builder produces the handoff package in step 1 only. The workflow then distills the reusable prompt and conventions (step 2), re-establishes them as durable project context (step 3), and reviews the package for gaps (step 4) so the next session starts warm and stays warm.
The package captures the project's current state, decisions made, open items, plus the distilled working prompt, conventions, and preferences worth carrying forward. Step 4's completeness review checks it for missing assumptions, constraints, files, risks, and next actions, giving a ready or not-ready verdict before anyone relies on it.
NewPrompt gives you the prompts, tool links, and step order; you run each prompt in your own AI tool. Capture the state with the Context Handoff Builder, distill the essentials, re-establish them as durable context the new session reads first, then review for gaps. You own the review and verdict.
The AI doesn't remember your project automatically forever — a handoff prepares a package to warm a new session, but the new session only stays informed if that context is re-established as standing input it reads first. Step 3 does exactly this, and step 4's gap check catches what would otherwise get dropped.
For a same-project new chat, step 1 may be enough. Steps 2 and 3 matter when the handoff has to persist — across many future sessions, or to someone else. Step 4, the completeness check, is worth it whenever someone else will rely on the handoff.
Starting a fresh AI chat shouldn't mean re-explaining the whole project. Here's how to turn the important parts of one conversation into a clean handoff you paste into the next.
Over a long project, AI keeps reopening decisions you already settled — re-suggesting the option you rejected, softening a constraint, renaming things. Here's how to turn those decisions into a block the model reads before it works.
You paste last month's project context and tell AI to continue — but half of it is superseded: the direction changed, a decision reversed, a deadline passed. The model treats the stale parts as current and aims the output wrong. Here's how to audit context before it misleads AI.
A long AI chat ends with a dozen suggestions — some you accepted, some you rejected, some already done — and "list the to-dos" turns all of them into pending tasks. Here's how to extract action items sorted by state, tied to evidence, with finished work kept out of the queue.
Get up to speed on an unfamiliar codebase in an afternoon — ground the AI in the project, have it explain the hard parts, and keep what you learn.
Turn a prompt that worked once into one you can reuse — pull the winning prompt out of the chat, mark the parts that change as variables, and lock it into a clean template.