ChatGPT Projects: How to Use Them (and Why You Should)

If your ChatGPT history is a wall of one-off chats you can never find again, ChatGPT Projects is the fix. A Project is a workspace that keeps the files, the chats and the standing instructions for one piece of work in a single place, so every conversation inside it already has the context. You stop re-explaining your situation at the top of every chat, and you stop losing good threads in the scroll.

Here's what ChatGPT Projects are, how to set one up, what to put in it, and how to use them well. It maps closely to Claude Projects, so if you use both tools the habits carry straight across.

What a ChatGPT Project is

A Project is a self-contained workspace inside ChatGPT. OpenAI describes Projects as "smart workspaces that keep everything related to a long-running effort in one place." Three things live in each one: The chats you have inside it, the files you upload as reference, and a set of instructions that shape how ChatGPT behaves there.

The value is standing context. A normal chat starts cold, so you supply the background every time. A Project carries the background for you, so every conversation in it already knows your material and your rules. It's the same instinct behind ChatGPT Work's projects, where you point the agent at folders of files, scaled down to the everyday chat you do all day.

Two features make a Project more than a folder. It holds files ChatGPT can draw on across every chat inside it, and it holds instructions that apply to all of them. Get both in place and a new chat in the Project needs only the task itself.

How to set up a ChatGPT Project

Creating one takes a minute. Click New project in the sidebar, give it a name, and pick an icon and colour so it's easy to spot later. Then, before you start chatting, add the context that makes it useful: The files, the instructions, and any connected apps the work needs.

That setup is the whole point. An empty Project is just a labelled folder; a Project with the right files and a clear set of instructions is a workspace that already understands the job. The two steps below are where that value comes from.

Adding project files

Files are how you hand ChatGPT your material once instead of pasting it into every chat. You can upload PDFs, spreadsheets, documents, images or plain text snippets, and ChatGPT draws on them for context across every conversation in the Project.

What goes in depends on the work. A content Project might hold your style guide, a few strong past pieces and your product details. A client Project might hold the brief, the contract and the notes so far. A research Project might hold the source reports you keep referring back to. The test is simple: If it's something you'd otherwise paste in or re-explain, it belongs in the Project files.

There's a capacity limit worth knowing, and it scales with your plan. On the free plan you get 5 files per Project; on Go and Plus, 25; and on Edu, Pro, Business and Enterprise, 40. So the free tier is enough to try the habit, and a paid plan gives a Project room to hold a proper reference set. OpenAI's own walkthrough of working with files is a good companion if you're loading a lot in.

Writing project instructions

Files are what ChatGPT knows; instructions are how it behaves. In the Project settings you set custom instructions that apply to every chat in that workspace, and, usefully, they override your global custom instructions. So a Project can have its own voice and rules without you touching your account-wide settings.

Good instructions are specific. OpenAI's own example gives the flavour: "Act like my marketing mentor. Be concise. Use bullet points. Ask clarifying questions." That gives ChatGPT something concrete to act on, where "be helpful" gives it nothing. Treat it like briefing a capable new hire, the clearer the standing brief, the less you correct later. If you catch yourself giving the same steer in chat after chat, that steer belongs in the instructions.

Keep a Project's context clean

One feature quietly makes Projects better for focused work: Memory. A Project can run in a project-only memory mode, where ChatGPT draws only on the conversations and files inside that Project and ignores your other chats. Shared Projects always work this way. That's what stops context from one piece of work bleeding into another, and it's why a Project stays on-topic in a way a long general chat never does. It's the same principle we cover in how ChatGPT works: The tighter and more relevant the context, the better the output.

The practical upshot is that you should give a distinct piece of work its own Project rather than piling everything into one. Separate Projects keep each context clean, and clean context is what keeps the answers sharp.

Projects vs a custom GPT

People often ask which to use, and the line is clean. A custom GPT is something you build to share, a packaged assistant with a fixed set of instructions, often published for other people to use. A Project is your own private workspace for an ongoing piece of work, where the files and instructions serve you and the chats stay together.

Use a custom GPT when you want to hand the same repeatable assistant to a team or the public. Use a Project when you want a focused space for work you keep returning to, with its own files and its own rules. For most day-to-day work, a Project is the one you'll reach for, because it takes a minute to set up and needs nothing published or maintained for anyone else.

How to actually use Projects well

The Projects that earn their keep are built around something ongoing, not a one-off question. If you return to the same kind of work often, a client, a product, a recurring report, a body of research, that's a Project. A thing you'll do once is fine as a normal chat.

Start small and let it grow. Create the Project, load the handful of files that matter most, write a short set of instructions, and use it. When you notice ChatGPT missing something, add the file that would have helped. When you catch yourself repeating a correction, move it into the instructions. A Project gets better by pruning and adding as you go, not by being perfect on day one.

Keep the files current, too. Out-of-date material quietly steers ChatGPT wrong, so swap the old version out when you add the new one. And one gotcha to file away: A scheduled task created inside a Project can't reach the Project's uploaded files, so plan any automation around that limit rather than being surprised by it.

Projects are one piece of setting ChatGPT up to do real work for you, alongside connected apps and instructions. Getting that whole setup right, so you delegate instead of re-brief, is what our Cowork training programme teaches, function by function, on the tools your team already runs. If you want to be the person who actually makes AI work rather than the one still pasting context into every chat, that's where to start.

Common questions

What are ChatGPT Projects? They're self-contained workspaces that keep the chats, files and instructions for one piece of work in a single place. Every conversation inside a Project shares that context, so ChatGPT always has your material and follows your rules without you re-supplying them.

How do I create a ChatGPT Project? Click New project in the sidebar, give it a name, icon and colour, then add your files and set the project instructions before you start chatting. Those two steps are what make the workspace useful.

Are ChatGPT Projects free? Yes, Projects are available on the free plan and every paid plan. The difference is capacity: Free allows 5 files per Project, Go and Plus 25, and Edu, Pro, Business and Enterprise 40.

What's the difference between project instructions and custom instructions? Project instructions apply only inside that Project and override your global custom instructions there. So a Project can have its own voice and rules without changing your account-wide settings.

Do ChatGPT Projects keep my work separate? Yes. A Project can use project-only memory, where ChatGPT draws only on that Project's chats and files and ignores your other conversations. Shared Projects always work this way, which keeps each piece of work on-topic.