# UseMyContext glossary > This is the markdown version of https://usemycontext.ai/glossary, generated from the same source as that page. The HTML page is canonical. > > Plain-English definitions of the words you meet on UseMyContext.ai - context layer, MCP, portable context, structural blindness, compiled profile, composite, @handle, pending suggestion and more. The words you meet on UseMyContext.ai, each defined in a sentence or two, in the same terms the rest of the site and the [docs](https://usemycontext.ai/docs/) use. ## What is UseMyContext? **UseMyContext** (usemycontext.ai) is a private personal context layer any AI assistant can read over MCP. You write down who you are once, connect Claude, ChatGPT, Gemini or Perplexity, and stop re-introducing yourself. It is a service, and it is unrelated to React's `useContext` hook. ## What is a context layer? A **context layer** is one place that holds a working picture of you: who you are, what you are working on, and how you like things done. Instead of re-explaining yourself at the start of every chat, your assistant reads that picture from the layer. A working context layer is compiled, gives the model scoped reads, only ever accepts proposed writes, and is owned and audited. More: [What is a context layer?](https://usemycontext.ai/blog/what-is-a-context-layer) ## What is MCP? **MCP**, the Model Context Protocol, is an open standard that lets an AI assistant read from a source you choose and approve. Once your assistant supports MCP, it can plug into UseMyContext the same way it plugs into anything else, with no custom wiring. Claude, ChatGPT, Gemini and Perplexity all speak MCP, so you set your context up once and any of them can read it. The UseMyContext server is at `https://mcp.usemycontext.ai/mcp`. ## What is portable context? **Portable context** is context you own that travels with you to every AI you use, instead of being trapped inside one app. The memory built into ChatGPT or Claude stays inside that one product; your UseMyContext context is the same in every assistant you connect. ## What is the difference between AI memory and context? **AI memory** is what one AI happens to remember about you. It is trapped inside that one app, you can't see it, move it, or correct it, and it is different in every assistant you use. Context is what you choose to share, owned by you: one place, portable to every AI, visible, editable, and revocable in one click. ## What is human-curated context? **Human-curated context** is context composed and approved by you, not scraped from your chats. Every word of your UseMyContext profile is yours: a connected AI can only propose an update, which waits as a pending suggestion until you accept it. As more of the web becomes machine-generated, a record of you written by you is the primary source. ## What is structural blindness? **Structural blindness** means the public-facing surface that talks to your AI declares no data bindings, so it holds no readable file content of yours. Your data lives on an isolated, per-user data plane behind it. Every plan, including Free, gets the same structural-blindness guarantee. ## What is a compiled profile? Your **compiled profile** is the curated summary of you that a connected AI reads first, built only from facts you accepted. Open a fresh chat and it already knows your role, your projects and your tone. The `profile` tool reads it. ## What is a composite? A **composite** is the curated, shareable form of a profile: what a person you share with, your team, or a public @handle page sees. Your raw facts and files are never part of it, so sharing a composite never hands over your files. ## What is a fact? A **fact** is one thing about you that you accepted into a profile: your role, a project, a preference. Facts come from the interview, from text you paste in, or from a suggestion you approved, and your compiled profile is built from them. ## What is a project? A **project** is one separate profile in your account, with its own facts, files and compiled profile, so work and personal, or each client, stay apart. Free includes 2 projects and Premium includes 10. Your AI can ask which profile a chat is reading with the `list_profiles` tool. ## What is an @handle? An **@handle** is the name a profile is addressed by, like @yourprofile. You claim one for each project. Making a project private or deleting it does not free its handle - it stays reserved to you, so nobody can pose as you under a name people already know. Deleting your whole account releases it. ## What does Private, Specific people and Public mean? **Visibility** is who can see a profile. Private means only you and your connected AIs. Specific people means a read-only invite, one direction, to the people you choose. Public means anyone with the link can read it, and public profiles are listed in the [public directory](https://usemycontext.ai/directory). You can flip back to Private anytime and the link dies instantly. ## What is a public profile page? A **public profile page** is the page at `usemycontext.ai/@yourhandle` for a profile its owner set to Public. The same curated composite is readable by any AI as plain markdown at `usemycontext.ai/@yourhandle.md`, with nothing to install. Private profiles are never served this way. ## What is a pending suggestion? A **pending suggestion** is an update a connected AI proposed with the `suggest_update` tool. It waits on your Profile page until you accept or ignore it. Nothing an AI does ever edits your profile or files directly, and nothing is ever written back to your own drive. ## What is a shared context? A **shared context** is a profile someone deliberately shared with you. Your AI reads it with the `shared_context` tool. Only the curated composite is shared, never the owner's raw facts or files, and the owner can revoke it at any time. ## What is a team context? A **team context** is one shared context for a whole team on the Team plan ($100/month, up to 20 seats for now). Teammates' AIs read each other's up-to-date profiles, built from the composite each member chooses to share into the team. More: [What is UseMyContext Team?](https://usemycontext.ai/blog/what-is-team) ## What is one-click revoke? **One-click revoke** is how you cut an AI's access. From the Connect page in the app you can revoke any single client, or disconnect all clients at once. Revocation is enforced on the server: the token dies immediately, everywhere. ## What is the session-starter file? The **session-starter file** is a local file with your curated context in it, written by one command: `npx usemycontext pull`. File-reading AI tools such as Claude Code hooks, agent frameworks and CLIs load it at startup. Full guide: [the session-starter CLI](https://usemycontext.ai/docs/cli-pull) ## What are the thirteen tools? The **MCP tools** are the thirteen actions a connected AI can take: twelve reads and one write. The reads cover your compiled profile, your files, cited passages from your documents, exact counts over a table, and contexts shared with you; the one write, `suggest_update`, only ever files a pending suggestion. Every access leaves a record. Full list: [The thirteen tools](https://usemycontext.ai/docs/tools) ## What is the demo account? The **demo account** is a shared, read-only account anyone can try without signing up: sign in as `demo@usemycontext.ai` with the code `424242`. It holds a full profile, real working files and a shared context, so every tool returns something real. More: [Try UseMyContext in 30 seconds](https://usemycontext.ai/blog/try-usemycontext-in-30-seconds) ## What is the UseMyContext SDK? The **UseMyContext SDK** is the `usemycontext` package on npm: a typed client and a React hook that drop a user's compiled context into any AI app you build over MCP, while the service stays structurally blind to the content. Install it with `npm i usemycontext`.