TL;DR: UseMyContext.ai is published in the ChatGPT plugin directory. Open the listing, click "Install plugin", sign in, and ChatGPT can read your compiled profile and files - the same account and the same context Claude reads. One habit matters: tag @UseMyContext in each message that should use your context, because untagged replies answer from ChatGPT's own memory, not your profile.
What is the ChatGPT plugin directory listing?
The ChatGPT plugin directory is OpenAI's in-product catalog of plugins you can add to ChatGPT. UseMyContext.ai has a listing there, so instead of pasting a server URL by hand you open the listing and click "Install plugin". The listing points at our one MCP server, https://mcp.usemycontext.ai/mcp, and the connection it creates is the same account and the same profile you would get connecting any other way - including from Claude. Your context does not fork per client; every AI you connect reads the same compiled profile.
How do you install it?
Two ways, same result. The direct way: open the listing link and click "Install plugin". From inside ChatGPT: open Settings, then Plugins, then "Browse plugins", search for usemycontext, open the UseMyContext.ai result and click "Install plugin". Either way, ChatGPT asks you to sign in with your email and a code, and then the plugin is available in any chat. The full walkthrough for every client is in the Connect docs.
What can ChatGPT read once connected?
The connection exposes the same thirteen tools every client gets, and every one of them is a read except a single suggestion tool that never writes to your data. ChatGPT can read your compiled profile, list your profiles, and see your plan and storage metadata. It can list, search, and read the files you uploaded, answer a question from your documents with cited passages via askdocs, and run exact counts and sums over a spreadsheet with querytable. It can also read context other people deliberately shared with you. The one write, suggest_update, files a pending suggestion that waits in your app for your approval. The full tool-by-tool list is in the Claude Directory announcement; the tools are identical here. Two of the thirteen, search and fetch, exist specifically for ChatGPT - more on that below.
How do you make sure ChatGPT uses your context?
This is the honest difference from Claude, and it is worth one sentence of muscle memory. ChatGPT only consults a plugin when the message invokes it, so tag @UseMyContext (or pick it from the plus menu) in each message that should use your context. An untagged message falls back to ChatGPT's own memory - whatever the model happened to retain, not the profile you curated. If a reply seems to have forgotten who you are, check the tag before anything else.
A connection also reads one profile at a time - your active one by default. Facts on your other profiles stay out of scope until you ask for a profile by its @handle or switch profiles at usemycontext.ai.
Does it work with ChatGPT deep research?
Yes, by design. Deep research requires a plugin to speak its search-and-fetch contract, and search and fetch are read-only aliases of our file search and file read that return results in exactly that shape. Point a deep research run at a question your documents can answer and it can pull from your own files alongside the open web.
What if my ChatGPT does not show the directory?
ChatGPT's menu names vary by version and plan, so the plugin browser is not always where you expect. On paid plans you can add it by hand: open Settings, then Apps, open Advanced settings and turn on Developer mode, click "Create app", name it UseMyContext, paste https://mcp.usemycontext.ai/mcp as the Connection, and click Create to sign in. On older ChatGPT versions this lives under Settings, then Connectors. It reaches the same server and the same account as the one-click install.
How do you stay in control after installing?
Every client you connect appears on the Connect page in your UseMyContext account, and you can revoke any one of them, or disconnect everything at once, at any time. Reads are audited, the connection is read-only against your data, and the public-facing surface of the service is built so it physically cannot read your file content. The only write ChatGPT can ever make is a suggestion that waits for your approval.
FAQ
Does installing from the directory cost anything?
No. The free tier includes 2 projects and 20 MB of storage for your profile, facts, and file uploads, so installing from the directory costs nothing on our side. The manual Developer-mode route is the one path our docs gate to a paid ChatGPT plan.
Is it the same account as connecting Claude?
Yes. The ChatGPT listing, the Claude Directory listing, and a manual custom connector all point at the same MCP server, mcp.usemycontext.ai/mcp, and read the same compiled profile. There is one account and one server behind every door, which is the point: build your context once, and every AI you use reads it.
Why does ChatGPT sometimes answer without my context?
ChatGPT answers without your context when the message did not invoke the plugin. ChatGPT only consults UseMyContext when a message tags it, so type @UseMyContext (or pick it from the plus menu) in any message that should read your profile.
Can ChatGPT change my files or profile through this connection?
No. The connection is read-only against your data. The one write tool, suggest_update, files a pending suggestion that waits in your app for your review; nothing is written without your approval.
How do I disconnect?
Open the Connect page in UseMyContext and revoke the connection. Access ends immediately, including any session already open. Whatever you change on ChatGPT's side, revoking here is the guarantee.