# What you may do with the data

> The licence in plain terms: build what you like on the data, but don't become a source of the dataset itself.

Source: https://hyperjobs.io/help/data-licensing

The short version: **you may build almost anything on the data, including things
you charge for. What you may not do is become a source of the dataset itself.**

This page is a summary. The [Terms of Service](/legal/terms) are the agreement,
and they win wherever this page and they disagree.

## What you may do

While your account is in good standing, the licence lets you use, copy, store,
modify and combine the data for your own internal business purposes and inside
your own products:

- **Store it and keep it synced.** Mirror it locally; [sync the feed](/docs/guides/feed-sync).
- **Enrich and transform it**, and combine it with your own data.
- **Show it to your users inside your product** — including a product you charge
  for.
- **Build analyses, models, dashboards and reports** on top of it.
- **Show it to your clients** as part of a service you deliver.

<Check>
  The line is not "you may not make money from this". Commercial use is the point.
  A job board, a market-intelligence dashboard, a recruiting tool, a model trained
  on hiring signal — all fine.
</Check>

## What you may not do

<Warning title="Don't be a pipe">
  Do not sell, licence, publish or otherwise supply the postings we return — in
  bulk, as a feed, as a file, as a database, or through an API of your own — as a
  substitute for someone else buying access from us.

  Data surfaced inside a product that *does something with it* is fine. A pipe
  from us to a third party is not.
</Warning>

The restrictions in full are in [§6 of the terms](/legal/terms). In summary, you
may not:

- **Resell or redistribute the dataset as a dataset.**
- **Build a competing jobs dataset for resale.**
- **Systematically reconstruct the corpus** — enumerating or paging the API to
  extract the whole dataset, rather than retrieving what your use case needs.
- **Circumvent quotas or rate limits** — multiple accounts to multiply a quota,
  keys shared across organisations, IP rotation. If you need more volume,
  [buy more](/pricing) or ask us.
- **Remove or obscure attribution** where you're required to keep it, or
  misrepresent where the data came from.
- **Reverse-engineer** the API or its classifiers.

## Acceptable use

There's no separate acceptable-use policy — [§7 of the terms](/legal/terms) is
it. The parts people trip over:

- **No unlawful discrimination.** Don't use enrichment fields to screen or target
  individuals on protected characteristics.
- **No spam.** Company data in the dataset is not a licence to send unsolicited
  bulk mail in breach of marketing or privacy law.
- **No profiling individuals.** Don't use this to identify, track, or build a
  dossier on a person. It's a jobs dataset, not a people dataset — see
  [the privacy policy](/legal/privacy).
- **No security testing without written permission in advance.** To report a
  vulnerability you've found, write to <Email kind="security" />.

## The common questions

<AccordionGroup>
  <Accordion title="Can I show these jobs on my own job board?">
    Yes. That's a product doing something with the data, and it's squarely inside
    the licence — including if you charge for it or run ads against it.
  </Accordion>

  <Accordion title="Can I train a model on it?">
    Yes, for your own products and internal purposes. What you can't do is ship
    the *corpus* — a model that reproduces the dataset for others to consume is
    the pipe problem wearing a different hat.
  </Accordion>

  <Accordion title="Can I give a client a CSV of the results?">
    Depends on what it is. A deliverable inside a service you're providing — an
    analysis, a shortlist, a report — is fine. Handing over a bulk extract so the
    client has the data instead of buying it is not.

    If the CSV is the product, ask first.
  </Accordion>

  <Accordion title="Can I cache the data?">
    Yes, indefinitely, and keeping a local mirror is the intended pattern for
    anything at volume. [Syncing the feed](/docs/guides/feed-sync) is the cheap
    way. Note that mirroring makes deletions your problem — a `404` on a
    previously-seen id means the employer took it down.
  </Accordion>

  <Accordion title="What happens to stored data if I cancel?">
    The licence runs while your account is in good standing and your fees are
    paid. Cancelling doesn't reach into your systems, but it does end the licence
    you were relying on — so if you're building something with a long life on
    stored data, read [§5 of the terms](/legal/terms) and ask us if the answer
    isn't obvious for your case.
  </Accordion>

  <Accordion title="I want the data as files, not API calls.">
    Flat-file delivery exists but isn't self-serve, and it isn't the same licence
    conversation as the API. Write to <Email kind="hello" />. There's no bulk
    export endpoint, and paging the API to build one is the reconstruction
    restriction above.
  </Accordion>
</AccordionGroup>

## If you're not sure

Ask before you build, not after. Write to <Email kind="hello" /> with what you
have in mind. We'd much rather answer the question early than have the other
conversation later — and the answer is usually yes.

## Next

<CardGroup cols={2}>
  <Card title="Terms of Service" icon="Scale" href="/legal/terms">
    The agreement itself.
  </Card>
  <Card title="Where the data comes from" icon="Database" href="/help/data-sources">
    Sources, refresh cadence, expiry.
  </Card>
</CardGroup>