What Are ChatGPT Projects?
ChatGPT projects are named workspaces that group everything for one job in a single place: the chats about that job, the reference files it needs, and custom instructions that apply to every chat inside. You set the context once instead of re-explaining it in every conversation.
Think of a project as a context container rather than a folder. That is the point of chatgpt projects: a folder only stores things, while a project stores things and tells ChatGPT how to work with them — the audience, the tone, the assumptions and the format you want back. That matters in 2027, when one account can hold several ongoing jobs and repeating the same setup in every chat wastes time.
OpenAI launched Projects on 13 December 2024, first for Plus, Team and Pro subscribers. Projects are now available on every ChatGPT plan, including Free, with smaller limits on the Free plan. You can open ChatGPT at chatgpt.com and find Projects in the sidebar.
Table of Contents
Key Takeaways
- ChatGPT projects bundle chats, files and custom instructions for one ongoing job, so you set the context once instead of repeating it in every chat.
- Create a project from the sidebar, name it after the outcome, and write instructions covering role, tone, audience and output format.
- Add reference files such as PDFs and spreadsheets; since the April 2026 redesign, files live in the Library and are referenced from the project.
- Projects are on every plan including Free, with smaller file limits on Free — check the Help Center for current limits.
- A project is for one job with a beginning and an end — not a filing cabinet. It has no nested folders and no project-wide search.
How to Set Up chatgpt projects for a Real Job
Setting up a project takes a couple of minutes. The naming and the instructions are the parts worth slowing down for — they decide whether the project saves you time or becomes another cluttered list.
How do I create a ChatGPT Project?
Open ChatGPT and look in the sidebar for the option to create a new project. Give it a name, pick an icon and a colour so you can spot it at a glance, and it is ready. Name it after the outcome, not the topic: “Spring 2027 Marketing Course Plan” beats “Marketing”, because it tells you what finished looks like. Once it exists, add files, write the project instructions, or move existing chats into it — in any order, though instructions first is the sensible default.
How do project instructions work?
Project instructions are short directions that apply to every chat inside that project: the role ChatGPT should take, the tone, the audience, the assumptions, the output format, and any rules for that job.
Good instructions are specific. “Write clearly” helps nobody. “Write for adult learners who are new to marketing, use plain language, define any technical term the first time you use it, and end every lesson outline with three discussion questions” gives ChatGPT something it can actually follow. Keep instructions short — a handful of lines beats a page of contradictory rules. Our guide to ChatGPT prompts will help you shape instructions that get consistent results.
What files can I add to a project?
You can add PDFs, documents, spreadsheets, images and pasted text as reference material — a syllabus, a set of readings, last year’s plan, a spreadsheet of dates, or images you want described or compared.
One change worth knowing about: since the April 2026 redesign, files live in the Library and are referenced from the project rather than stored inside it. In practice, files are organised in one place and the project points to the ones it needs. Add only the files the job genuinely needs — five defining documents beat twenty loosely related ones.
A Worked Example: Planning a Course With ChatGPT Projects
The easiest way to understand projects is to build one for a real job. Here is a complete setup for planning a short course — the same pattern works for a grant application or any piece of work that stretches over several sessions.
Suppose you are planning a six-week evening course for adults, “Introduction to Digital Marketing”, starting in spring 2027. Before the project, each chat would begin with you re-explaining the audience, the level and the format. With a project, you explain it once.
Step 1: Create the project. From the sidebar, create a new project named after the outcome: “Spring 2027 Intro to Digital Marketing Course”. Pick an icon and colour you will recognise instantly.
Step 2: Write the project instructions. Something like this: “You are helping me plan a six-week evening course for adult learners new to digital marketing. Use plain language, define technical terms the first time you use them, and make each lesson outline fit a two-hour session with one practical exercise and three discussion questions.”
Step 3: Add the reference files. Upload the files the course depends on: session dates, the textbook chapter list, the budget spreadsheet, and notes from last year’s course. Leave out anything that belongs to a different job.
Step 4: Start one chat per distinct outcome. Open a new chat for the course outline, another for the first lesson plan, and another for the exercise sheets — one deliverable per thread.
Step 5: Check the output against your sources. Read the outline against your dates and chapters before you trust it. Projects make ChatGPT more consistent, not more accurate.
The same five steps fit a grant application: name the project after the grant and the deadline, write instructions covering the funder’s audience and word limits, add the funder’s guidelines and budget spreadsheet, and keep one chat per section.
ChatGPT Projects vs Plain Chats vs Folders: What Is the Difference?
The most common confusion about projects is treating them as folders with a new name. They are not. A folder organises; a project organises and instructs. That distinction decides which one to use.
| ChatGPT Projects | Plain chats | Folders (organising chats) | |
|---|---|---|---|
| What it holds | Chats, files and instructions for one job | One conversation at a time | A group of chats under one label |
| Instructions | Apply to every chat inside | You repeat the context in each new chat | No shared instructions — grouping only |
| Files | Available to all chats in the project | Files attached to one chat only | Files stay with their individual chats |
| Best for | One ongoing job, several sessions | One-off questions and quick tasks | Tidying a long chat list by topic |
| Memory | Can draw on chats within the project | General account memory only | No effect on memory |
| Sharing | Business, Enterprise and Edu plans | Not shareable as a workspace | Not shareable as a workspace |
If you only need to find an old chat again, organising your chats is enough — and knowing how to delete ChatGPT history helps keep the list manageable. If you need ChatGPT to remember the setup for a job across many sessions, that is what a project is for.
Are ChatGPT Projects Free?
Yes. Projects are available on every ChatGPT plan, including Free, so there is no separate charge. What changes between plans is how much you can put in a project — most importantly, the file limits.
As listed in OpenAI’s Help Center as checked in November 2026, the per-project file limits are 5 files on Free, 25 on Plus, and 40 on Pro, Business and Enterprise. Sources have conflicted on the exact caps, so treat these as the Help Center’s current listing, not a permanent promise — check the OpenAI Help Center article on Projects in ChatGPT before planning around a limit. OpenAI’s original announcement, Introducing ChatGPT Projects, gives the background on what the feature was built to do.
In practice, Free suits a small job with a handful of reference files; Plus and above suit a larger body of source material. Either way, the limit that bites first is usually not the file count but the lack of structure — with no nested folders, a project with dozens of chats still needs clear chat names and one fresh chat per outcome.
What are the limits of ChatGPT Projects?
Beyond the file caps: no nested folders or sub-projects, no bulk export of a whole project, and no project-scoped search. Projects are built for one ongoing job with a beginning and an end, not as a filing cabinet for everything you have ever asked ChatGPT — if a second, unrelated job creeps in, make a second project.
What Is Project Memory?
Project memory is a newer option that lets ChatGPT draw context from conversations within the same project. Instead of each chat starting cold apart from the instructions and files, the project can use what was discussed in its other chats when it answers — useful for a course plan that builds week by week, or a grant application where a decision made in one chat affects the wording in another. Existing projects stay on default memory unless you change them. Memory is no substitute for files: a decision buried in a chat is harder to find than the same decision written into your instructions or reference documents.
Can I Share a ChatGPT Project?
Business, Enterprise and Edu users can share projects with teammates, so a team works from the same instructions and files instead of each person rebuilding the setup. Personal memories are disabled for shared projects — nobody’s private conversation history should leak into a shared workspace.
Sharing makes the quality of your instructions matter more, not less. When you are the only reader, a vague instruction costs you a follow-up question. When a whole team inherits it, a vague instruction costs everyone one — so write shared instructions as if a new colleague will read them with no other explanation.
How Do I Move a Chat Into a Project? Can I Search Inside a Project?
To move an existing chat into a project, right-click the chat and choose Move to Project. The moved chat inherits the project’s instructions and file context, so a conversation you started before the project existed can still benefit from the setup.
Searching is a weak spot: there is no project-scoped search. The workaround is discipline — give each chat a clear name that says what it produced, keep one outcome per chat, and keep important decisions in the instructions or files. If you use ChatGPT’s focused modes alongside projects, our guide to ChatGPT Study Mode shows how one such mode changes the way answers are structured.
When Should I NOT Use a Project?
Projects are the wrong tool more often than people admit. Do not create one for a one-off question — a plain chat is faster. Do not create one as an archive; with no nested folders and no project search, a project stuffed with unrelated conversations is harder to use than the list it replaced. And do not create one for a job with no shared context — if every chat needs different instructions and different files, the project adds a layer without adding value.
Ask one question before creating a project: will I return to this job at least three times? If yes, and the sessions will share instructions and files, make the project. If no, use a plain chat. And if the job is really a library of tools rather than a single piece of work, browse our roundup of the best AI tools to find a tool built for that task.
For more new features and practical guides, see the Tech Insights section on Digital Discoverist — and the rest of our coverage at Digital Discoverist.
FAQ
Do ChatGPT Projects cost extra?
No. Projects are included on every ChatGPT plan, including Free. Only the limits, such as how many files each project can hold, change between plans.
What happens if I hit the file limit on the Free plan?
On Free, each project holds 5 files as listed in OpenAI’s Help Center as checked in November 2026. Keep only the files the job genuinely needs, and check the Help Center for the current limits.
Can I use one project for two different jobs?
You can, but you should not — projects are built for one ongoing job, and there are no nested folders to keep two jobs apart. Make a second project instead.
Where do my project files live?
Since the April 2026 redesign, files live in the Library and are referenced from the project rather than stored inside it.
3 thoughts on “ChatGPT Projects in 2027: What They Are and How to Use Them”