Your Playbook files were written on a particular day. Like most technology - and especially when it comes to AI - Claude makes changes. A lot. If what's on your screen doesn't match what's on the page, that isn't you doing it wrong. It usually means something changed or was updated. This page tells you what changed and what to do or look for instead. Most recent changes will be listed first.
Affects: projects you made before the new version reached your account
Upgrading is simple for most projects. Open your project and look for a card that says "Let Claude coordinate this project." Click Get started, then Upgrade project. Keep copies of your files on your computer first, because the upgrade can't be undone.
The full walkthrough, with pictures: How to Move Your Project to the New Claude Projects.
Affects: the Project-Based path, if the change has reached your account
Anthropic is rolling out a new version of Claude Projects. A Project used to work like a folder: you opened it and started a new chat inside it for each task. In the new version, a Project is one ongoing conversation. You ask for things as they come up, and when a piece of work needs its own space, Claude branches it into a thread and reports back as it goes.
It is reaching accounts in stages. Some people already have it in the Claude app, on the phone as well as the computer. Others don't yet, and Anthropic hasn't said when every account gets it. If yours hasn't changed, your Project works exactly the way your guide describes.
When it reaches you, you'll see a screen that says "A project is now a conversation with Claude." Tap Continue. Anthropic says project instructions and uploaded files keep working, so your setup carries over.
What to do when you see it. Your guides and almost every prompt say to start a new chat inside your Project or workspace. In the new version, skip that step: open your Project and type the prompt into the conversation, for example "Run P1-01." Don't go looking for a new chat outside your Project, because Claude can't see your files there. Claude starts a thread for it. When the thread needs an answer from you (most prompts start by asking questions), click View thread and answer in that thread's message box. Each thread has its own model and effort setting, so for the heavy prompts (P1-01, P2-01 and any pipeline audit) check that the thread is set to the most capable model and a high effort.
The v1.1 Playbook files are written for the new version, so they already describe threads.
Source: What are projects? (Claude Help Center) and Anthropic's announcement, 17 September 2026
Affects: any screenshot showing the model picker
Your guides deliberately never tell you to pick a model by name, because Anthropic renames and replaces them faster than a PDF can keep up. They tell you to pick the most capable one available. A screenshot, though, has to show something, so it will show whatever was current the day it was taken.
What to do. Open the model picker and read the descriptions rather than the names. You want the one described as being for your hardest or most complex work, and the Playbook runs every thread on it at High effort, which gave noticeably better results than Medium in my testing. My current pick is on the Help page. If you ever want a different model or effort for one task, just tell Claude in your message.
That's everything. If something in your guide doesn't match your screen and it isn't listed here, email alva@getseenplaybook.com and I'll look into it and see if it needs to be added to this page.