The ChatGPT apps, projects, and memory
The same ChatGPT account follows you across three surfaces (web, desktop, and mobile), but each one is tuned for a different moment. Underneath, the features that actually change your workflow, Projects and memory, behave the same everywhere. This lesson tours the surfaces fast, then goes deep on the two things that turn ChatGPT from a chat box into a working environment.
The three apps, and what each is actually for
The web app (chatgpt.com) is the full surface. Everything ships here first: Projects, Canvas, Advanced Data Analysis, apps (formerly called connectors), scheduled tasks, the GPT Store. If a feature is missing on another surface, check the web first.
The desktop app (macOS and Windows) adds OS-level reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition →. On macOS it can see other app windows when you grant permission, so you can ask it about the spreadsheet or code editor in front of you without copy-pasting. A global keyboard shortcut opens a quick launcher, which is the fastest way to start a chat without breaking focus. Companion tools like Codex (OpenAI's coding agent) fit most naturally next to a desktop dev environment.
The mobile app (iOS and Android) is built for input you can't easily type: voice mode for hands-free conversation, the camera for "what is this" questions, and photo uploads from your library. Voice is the standout here.
The mental model: start serious work on web, capture and dictate on mobile, and use desktop when you want ChatGPT aware of what's already on your screen. Your chats, Projects, and memory sync across all three.
Projects: a scoped workspace, not just a folder
A Project is a container that bundles three things: **a set of chats, uploaded files, and custom instructions that apply *only* inside that container**. Think of it as a sandbox with its own brief and its own reference shelf.
Why this matters: your global custom instructions and global memory are good defaults, but they apply to everything. A Project lets you set instructions and load files for one body of work without cluttering the rest of your account.
What lives inside a Project
- Project instructions. A scoped system promptsystem promptThe hidden set of instructions that defines how an AI assistant behaves before any user types a question: its role, tone, limits and rules.View full definition →. Anything you write here applies to every chat in the Project.
- Project files. Documents you upload once and reference from any chat in the Project. ChatGPT retrieves from these automatically when relevant.
- The chats themselves. Every conversation started inside the Project stays grouped there.
Memory in Projects depends on a choice you make when you create the Project. In the default mode, chats inside the Project can use your global memory, and ChatGPT also draws on other chats in the same Project. If you pick project-only memory, the Project can only reference its own chats and won't pull in memories from outside it. You can't switch this after creation, so decide up front. For client work with confidential context, project-only is usually the right pick.
Projects can also be shared. Invite collaborators and they see the same files, instructions, and chats. Limits and plan availability change often, so check OpenAI's Projects help article for current specifics.
A concrete Project: the "Northwind Retail" client
Say you consult for a client, Northwind Retail, on a quarterly basis. Without a Project, every new chat starts cold: you re-paste their brand voice, re-explain their product lines, re-upload the same brief. Tedious and error-prone.
Create a Project called Northwind Retail and choose project-only memory. Set the instructions:
You are my analyst and copywriter for Northwind Retail, a mid-market
outdoor-gear retailer. Brand voice: practical, warm, no hype, no exclamation
marks. Always assume a North American audience. When I ask for analysis,
show your reasoning and cite which uploaded file you used. Currency: USD.
Default to plain language a store manager can act on.Then upload the reference shelf as Project files:
northwind-brand-guide.pdfq3-sales-by-sku.csv2025-marketing-calendar.xlsxstore-locations.csv
Now every chat in this Project already knows the voice, the audience, and the data. You open a new chat and ask: *"Draft three email subject lines for the fall jacket promo, then tell me which SKUs from the sales file are trending down."* ChatGPT applies the brand voice from instructions and pulls numbers from q3-sales-by-sku.csv without you uploading anything again.
Stacking project files with advanced data analysis
Project files get much more useful when combined with Advanced Data Analysis (Code Interpreter), the sandboxed Python environment inside ChatGPT. Because your CSV is already on the Project shelf, you can say:
*"Using the sales file, chart weekly revenue by category and flag any category that dropped more than 15% week over week."*
ChatGPT writes and runs real Python against the file, returns the chart, and explains the dips, all within the Project's brand-aware context. No re-upload, no re-promptingpromptingPrompt engineering is the practice of designing and refining text inputs to guide large language models toward accurate, relevant, and reliable outputs.View full definition → the voice.
When to split into multiple Projects
One Project per ongoing client or per major workstream. If you also write Northwind's internal HR docs, that's a *different* tone and a different file set: make a second Project. Separate Projects keep the brand-voice instructions from bleeding between unrelated work.
Memory: what ChatGPT carries between chats
Memory is ChatGPT's ability to remember details across separate conversations so you don't repeat yourself. It comes in two parts.
Saved memories are specific facts ChatGPT chooses to store, like "User is a consultant" or "User prefers metric units." You can see them in Settings → Personalization → Memory, and you can delete any one of them.
Chat history reference is broader: ChatGPT can draw on the themes and context of your past conversations to make new answers more relevant, even without an explicit saved memory.
Both are toggles. Both live under Personalization. The full, current rules are in OpenAI's Memory FAQ.
How memory helps in the Northwind example
Outside the Project, in your everyday chats, memory learns that you're a consultant who writes for retail clients, prefers concise drafts, and works in USD. So when you start a quick unrelated chat to outline a proposal, ChatGPT already matches your style. Less re-explaining, more continuity.
Inside the Northwind Project, the *Project instructions* do this job instead, scoped tightly to that client. Because you chose project-only memory, Northwind details stay inside that Project. The division of labor: global memory for your durable personal defaults, Project instructions and project-only memory for client-specific context.
When to turn memory off
Turn it off (or use a Temporary Chat) when:
- You handle multiple clients with conflicting context. You don't want a stray fact about Northwind surfacing while you draft for a competitor. Project-only memory handles most of this, but for ad hoc chats outside any Project, a Temporary Chat is the safe move.
- You're doing sensitive or one-off work. A Temporary Chat starts a session that doesn't use memory, doesn't update memory, and doesn't appear in history.
- Memory has gone stale. If ChatGPT keeps assuming an outdated fact (an old job title, a past project), open the memory list and delete the offending entry rather than fighting it in conversation.
You can ask ChatGPT directly: *"What do you remember about me?"* and it will list its saved memories. You can also say *"Forget that I work at Northwind"* and it will drop that specific memory.
ChatGPT Projects and Memory Explained
Knowledge check
1. According to the lesson's mental model, when should you prefer the desktop app over the other surfaces?
2. What best distinguishes a Project from simply using ChatGPT with your global custom instructions?
3. How do Project files function within a Project according to the lesson?
4. Select ALL statements that correctly describe how the three ChatGPT surfaces relate to each other.
Select all the correct answers.
5. Select ALL of the elements that a Project bundles together as described in the lesson.
Select all the correct answers.
How this connects to the rest of the ecosystem
Projects and memory are the *personal and team* layer. They sit alongside the more programmable parts of the OpenAI ecosystem, and knowing the boundaries keeps you from reaching for the wrong tool.
Custom GPTs are packaged assistants with their own instructions, knowledge files, and optionally GPT Actions (calls out to external APIs). A shared Project is for collaborating on one ongoing body of work with a few people. A Custom GPT is a reusable tool you build once and hand to many people, or publish to the GPT Store. If five teammates are working on the Northwind account together, share the Project. If fifty sales reps need a Northwind brand-voice assistant, build a Custom GPT.
Apps in ChatGPT (what OpenAI used to call connectors) link ChatGPT to external sources and services like Google Drive, SharePoint, or GitHub so it can search them live and, in some cases, take actions. Project files are a fixed shelf you upload. Apps are a live pipepipeAll active sales opportunities across the stages of the sales process, together with their combined potential value and probability of closing.View full definition → to a system of record.
The ChatGPT agent and scheduled tasks add autonomy: the agent can browse and act across steps to complete a task, and scheduled tasks let you say *"every Monday at 8am, summarize last week's Northwind sales file and email me the top three movers."*
When you outgrow the app: the APIAPIApplication Programming Interface: a standardised interface that lets applications communicate and exchange data without knowing each other's internal workings.View full definition →
When you need this repeatable and embedded in software, you move from the app to the OpenAI API. The recommended entry point is the Responses API, which unifies model calls with built-in tools. (The older Chat Completions API still works and is widely supported.) For reliable downstream parsing, use structured outputs to force the model to return JSON that matches a schemaschemaA schema is the formal blueprint that defines how data is structured, named, typed, and related within a database, file, or message.View full definition → you define.
Here's the Northwind "flag declining SKUs" task as a small, schema-constrained API call. Swap in whichever current model fits your cost and speed needs from the models page:
from openai import OpenAI
client = OpenAI()
response = client.responses.create(
model="gpt-5",
input="From the sales summary, list SKUs with a week-over-week revenue drop over 15%.",
text={
"format": {
"type": "json_schema",
"name": "declining_skus",
"schema": {
"type": "object",
"properties": {
"skus": {
"type": "array",
"items": {
"type": "object",
"properties": {
"sku": {"type": "string"},
"wow_change_pct": {"type": "number"},
},
"required": ["sku", "wow_change_pct"],
"additionalProperties": False,
},
}
},
"required": ["skus"],
"additionalProperties": False,
},
}
},
)
print(response.output_text)That structured result drops straight into a dashboard or an email job, no manual parsing. From here, function calling lets the model trigger your own tools, and the Agents SDK orchestrates multi-step agents in code. Start at the OpenAI API docs when you cross this line.
The progression: Projects and memory for your own and small-team work, Custom GPTs and apps for packaged, widely shared assistants, the API and Agents SDK for production software.
Key Takeaways
- Use one Project per ongoing client or workstream. Load its files once and write scoped instructions; every chat inside inherits the brief.
- Pick the Project's memory mode at creation. Default mode shares global memory; project-only keeps context inside the Project, and you can't change it later.
- Combine Project files with Advanced Data Analysis to run real Python against your uploaded data without re-uploading or re-prompting the brand voice.
- Let global memory hold durable personal defaults (your role, style, units) and let Project instructions hold client-specific context.
- Reach for a Temporary Chat for sensitive or one-off work, and prune stale facts in Settings → Personalization → Memory instead of arguing with the model.
- Share a Project to collaborate, build a Custom GPT to distribute, connect apps for live data, and move to the Responses API with structured outputs for production.
Frequently asked questions
What is the difference between a ChatGPT Project and a regular chat?
A Project is a container that bundles chats, uploaded files, and custom instructions that apply only inside it. A regular chat starts cold each time and relies on your global instructions and memory. Inside a Project, every new conversation already knows the brief, the tone, and the reference files you loaded once.
Should I use the ChatGPT web, desktop, or mobile app?
Start serious work on the web app (chatgpt.com), because every feature ships there first: Projects, Canvas, Advanced Data Analysis, apps, scheduled tasks, the GPT Store. The desktop app on macOS can see other app windows with permission, so use it when you want ChatGPT aware of the spreadsheet or editor in front of you. Mobile is for voice mode, camera questions, and photo uploads. Chats, Projects, and memory sync across all three.
How does ChatGPT memory actually work?
ChatGPT memory has two parts. Saved memories are specific facts it stores, like "User is a consultant" or "User prefers metric units", visible and deletable one by one in Settings → Personalization → Memory. Chat history reference is broader: ChatGPT draws on themes from past conversations even without an explicit saved fact. Both are separate toggles under Personalization.
Can I stop ChatGPT from mixing context between two competing clients?
Yes. When you create a Project, choose project-only memory: that Project can then reference only its own chats and won't pull in memories from outside. This choice is locked at creation and cannot be changed afterwards. For one-off chats outside any Project, use a Temporary Chat, which neither uses nor updates memory and leaves no history.
When should I build a Custom GPT instead of sharing a ChatGPT Project?
Share a Project when a handful of people collaborate on one ongoing body of work and need the same files, instructions, and chat thread. Build a Custom GPT when you want a reusable assistant handed to many people or published to the GPT Store. Rule of thumb: five teammates on one account, share the Project; fifty sales reps needing the same brand-voice assistant, build a Custom GPT.
How do I analyze a CSV uploaded to a ChatGPT Project without re-uploading it each time?
Upload the CSV once as a Project file, then call Advanced Data Analysis (Code Interpreter) from any chat in that Project. Ask something like "using the sales file, chart weekly revenue by category and flag any category that dropped more than 15% week over week" and ChatGPT writes and runs real Python against the file. The Project instructions still apply, so the output keeps the tone and assumptions you set.
What to do, from this lesson
These actions are compiled in the role's Playbook.
- Run one Project per client, scoping files and instructions to it
- Reserve global memory for durable personal defaults, not project specifics
Related articles
Recent articles from the blog that build on this lesson.