> ## Documentation Index
> Fetch the complete documentation index at: https://docs.oncortex.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Quickstart

> From new account to an agent that remembers, in about ten minutes

By the end of this page you will have a Cortex account, something filed in your own brain, and an AI agent that can recall it.

<Steps>
  <Step title="Create your account">
    Go to [app.oncortex.ai](https://app.oncortex.ai) and register with your email address, or sign in with Google. If you register with email, confirm your address from the email we send you, then sign in and complete your profile: your name and a workspace name (your company name, or just your own).
  </Step>

  <Step title="Meet your brain">
    Open **Brains**. You already have one, called **My Work Brain**. It was created with your account, and its member list starts with exactly one person: you. Nobody else can see into a brain they are not a member of, so it is yours alone unless you one day choose to share it. This is where everything you capture lands unless you deliberately send it somewhere else.
  </Step>

  <Step title="Capture your first item">
    Open your brain and go to its **Inbox**. Drag in a file: meeting notes, a PDF, a Word document, anything with words in it. It appears in the inbox as a new item, with its text extracted and ready to file.

    Nothing goes straight into the brain. The inbox is the tray on top of the filing cabinet: everything waits there until it is filed.
  </Step>

  <Step title="Watch it get filed">
    Straightforward items are filed for you: a page is created in the right part of the brain, written in plain language, with a note of where the information came from. Anything Cortex is not confident about is parked in the review queue instead, with a suggested filing and the reasoning, waiting for your yes or no.

    You can also file items yourself, or have your connected agent do it, which brings us to the good part.
  </Step>

  <Step title="Connect your agent">
    Open your brain's **Connect** page. You will see your workspace's connection URL. Add it to your AI tool:

    * **Claude (web or desktop):** add it as a connector and approve the connection when Cortex asks.
    * **Claude Code:** copy the one-line `claude mcp add` command shown on the Connect page, run it, and approve in the browser window that opens.
    * **Anything else that speaks MCP:** paste the URL as a remote MCP server.

    The approval screen names your workspace and shows that the agent will act as you, so you always know exactly what you are granting. One connection covers every brain you can access, at the roles you actually hold.
  </Step>

  <Step title="Ask your agent to remember">
    In your agent, ask something like *"Search my brain for what I uploaded earlier"* or *"What do you know about me so far?"*. The agent searches your brain and answers with what it finds, citing the pages it read.

    From now on, anything you tell your agent to save, and anything you drop in the inbox, becomes part of what it remembers.
  </Step>
</Steps>

## Where next

<CardGroup cols={2}>
  <Card title="Core concepts" icon="brain" href="/core-concepts">
    Brain, page, inbox, skill, connection: the five nouns in two sentences each.
  </Card>

  <Card title="The inbox in depth" icon="inbox" href="/capture/the-inbox">
    Statuses, the review queue, and what happens to duplicates.
  </Card>
</CardGroup>