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

# Your first ten minutes

> Connect an assistant, teach it the rules, fill the brain, build the first resume. In that order — the order is the whole trick.

A new Hired workspace is empty, and an empty workspace does nothing. These four steps
are what turn it into something that answers questions about your career. They take
about ten minutes, and only the last one needs you to think.

<Note>
  If you are setting up an instance rather than using one, start at
  [Deploy it](/self-hosting/overview) and come back here once you can sign in.
</Note>

<Steps>
  <Step title="Connect an assistant" icon="plug">
    Sign in and open **Settings**. It opens on **Connections**, because this is the step
    everything else depends on. You already have one waiting.

    Hit **Set up**, pick the client you use, and the exact steps appear with your URL
    already filled in. Paste it where it asks. That is the whole procedure — no OAuth,
    no headers, no local process to keep running.

    Give each client its own connection rather than sharing one. That is what lets you
    disconnect a lost laptop later without breaking everything else.

    <Warning>
      A connection URL is a password. Anyone holding it can read and write everything in
      your workspace. It cannot reach anyone else's data, but it can reach all of yours —
      so never paste one into a chat, an issue or a screenshot. If one leaks,
      **Rotate** it.
    </Warning>

    [Per-client setup, in detail →](/connect)
  </Step>

  <Step title="Install the skills" icon="wand-magic-sparkles">
    A skill is a file that teaches an assistant how to behave *before* you ask it
    anything. Three ship with your instance, and they are the difference between an
    assistant that guesses at your career and one that knows the rules of this place.

    Install `hired` if you install only one — it carries the no-fabrication rule and the
    table of which tools overwrite rather than append.

    [Where each client wants them →](/skills)
  </Step>

  <Step title="Fill the brain" icon="brain">
    Do this before you build a resume. There is nothing to assemble a document out of
    otherwise.

    Your assistant already knows the workspace is empty — the briefing it got on connect
    says so, and tells it to ask you one question before anything else: do you have a
    resume or a LinkedIn export to paste, or would you rather talk it through? Either
    answer works, and it files what comes back by hand.

    So talk, or paste. Do not go looking for a form:

    > I worked at Vertex from March 2021 to last month as a senior platform engineer.
    > Here's everything I remember about it: …

    Length is a feature. The brain dump has no limit and no required structure — numbers,
    projects, stories, praise, the thing that went wrong. Nobody reads it directly; it is
    searched. The more raw material in there, the better every future resume is, and the
    fewer questions you have to answer twice.

    Twenty minutes of rambling per job is the right amount of effort. If you would rather
    do it in pieces, dump one job a day.

    [What goes in and how to get it in →](/guides/fill-the-brain)
  </Step>

  <Step title="Build the first resume" icon="file-lines">
    Now paste a real job posting and ask for the thing you actually want:

    > Tailor my resume to this posting, and tell me what I can't evidence.

    You get a saved document and — the useful half — a list of the posting's requirements
    with nothing behind them. That list tells you what to go and learn, or which jobs to
    skip.

    [How tailoring works →](/guides/tailor-a-resume)
  </Step>
</Steps>

## Then what

Once there is material in the brain and a document on file, the rest of the product is
just talking about what happened.

<CardGroup cols={2}>
  <Card title="Track something" icon="list-check" href="/guides/run-the-search">
    Paste a posting URL and it becomes a tracked application in one move — company,
    description, location, compensation and source all filled.
  </Card>

  <Card title="Log what happened" icon="clock-rotate-left" href="/guides/run-the-search">
    "I had a call with the Helios recruiter, the team is six people and they want
    someone to own billing." That is a timeline entry, said out loud.
  </Card>

  <Card title="Get the read" icon="gauge" href="/guides/weekly-review">
    `diagnose_search` names which step of the funnel is losing people, rather than
    handing you six numbers to interpret.
  </Card>

  <Card title="Publish it" icon="link" href="/guides/publish-and-export">
    A real PDF for the forms that want a file, an unlisted URL for the ones that want a
    link.
  </Card>
</CardGroup>

## The three mistakes people make first

<AccordionGroup>
  <Accordion title="Building a resume before filling the brain" icon="triangle-exclamation">
    There is nothing to build it from, so the assistant either produces something empty
    or starts inventing. Fill at least your current role first. If you already have a
    resume, paste its contents into a brain dump — it is a floor, not a ceiling.
  </Accordion>

  <Accordion title="Treating the brain dump like a resume" icon="triangle-exclamation">
    It is not a document, and nobody will read it as one. Bad grammar, half sentences and
    contradictory notes are all fine. What matters is that the specifics are in there:
    the number, the system name, the team size, what actually changed.
  </Accordion>

  <Accordion title="Editing a resume that is already attached to an application" icon="triangle-exclamation">
    That is the version you actually sent. Tailor into a copy instead — `duplicate_resume`
    exists for this, and an assistant that has the `hired` skill installed does it without
    being asked.
  </Accordion>
</AccordionGroup>