Sensitive Data Protection helps you discover, classify, and de-identify
sensitive data inside and outside Google Cloud. This page describes the services
that make up Sensitive Data Protection.

## When to use Sensitive Data Protection

The following are some common use cases:

- **Discover shadow data across enterprise cloud repositories**: Continuously scan and profile data assets across BigQuery, Cloud SQL, and Cloud Storage to locate uncatalogued sensitive data and evaluate organizational security posture.
- **Prevent sensitive data exposure in real-time applications**: Inspect and redact sensitive tokens (such as PII or credentials) from streaming API payloads, user forms, and LLM prompt inputs before storing the data.
- **De-identify analytical datasets while preserving utility**: Transform direct and indirect identifiers using pseudonymization, date shifting, or bucketing so research and business intelligence teams can analyze data without exposing plaintext PII.
- **Automate compliance auditing and regulatory reporting**: Detect and monitor protected health information (PHI) and financial records to satisfy compliance obligations under GDPR, HIPAA, and PCI DSS.
- **Stream security findings into incident response pipelines**: Send high-risk finding alerts to Pub/Sub and Security Command Center to trigger automated data tagging, access revocation, and security remediation workflows.

## Sensitive data discovery

The discovery service lets you generate profiles for your data across
an organization, folder, or project. Data profiles contain metrics and metadata
about your data assets and help you determine where [sensitive and high-risk
data](https://docs.cloud.google.com/sensitive-data-protection/docs/sensitivity-risk-calculation) reside. Sensitive Data Protection reports these metrics at various
levels of detail. For information about the types of data you can profile, see
[Supported resources](https://docs.cloud.google.com/sensitive-data-protection/docs/data-profiles.md.txt#supported-resources).

You use a *scan configuration* to specify the resource to scan, the
types of information ([*infoTypes*](https://docs.cloud.google.com/sensitive-data-protection/docs/infotypes-reference.md.txt)) to
look for, the profiling frequency, and the actions to take when profiling
is complete.

> [!TIP]
> **Tip:** Data profiling is useful if you want to scan large amounts of data at a high level. If you need to know the granular details, like the exact location of every instance of sensitive data, consider performing an [inspection](https://docs.cloud.google.com/sensitive-data-protection/docs/sensitive-data-protection-overview.md.txt#inspection) as well.

For more information about the discovery service, see [Data
profiles overview](https://docs.cloud.google.com/sensitive-data-protection/docs/data-profiles).

## Sensitive data inspection

The inspection service lets you perform a deep scan of an individual
resource to find instances of sensitive data. You specify the infoType that you
want to search for, and the inspection service generates a report about
every instance of data that matches that infoType. For example, the report tells
you how many credit card numbers are in a Cloud Storage bucket and the
exact location of each instance.

> [!TIP]
> **Tip:** An inspection is useful if you need
> detailed information about each instance of sensitive data stored in a resource,
> like a single
> BigQuery table. It is especially useful if you have
> unstructured data---like user-provided comments---that might have
> intermittent instances of personally identifiable information.
>
> If you
> need to perform automated scans of [multiple resources](https://docs.cloud.google.com/sensitive-data-protection/docs/data-profiles.md.txt#supported-resources) across projects, folders,
> or the entire organization, use the [discovery
> service](https://docs.cloud.google.com/sensitive-data-protection/docs/sensitive-data-protection-overview#discovery) to generate data profiles.

There are two ways to perform an inspection:

- Create an inspection or hybrid job through the Google Cloud console or through the Cloud Data Loss Prevention API of Sensitive Data Protection (DLP API).
- Send a [`content.inspect`](https://docs.cloud.google.com/sensitive-data-protection/docs/reference/rest/v2/projects.content/inspect.md.txt) request to the DLP API.

### Inspection through a job

You can configure inspection and hybrid jobs through the Google Cloud console
or through the Cloud Data Loss Prevention API. The results of inspection and hybrid jobs are
stored in Google Cloud.

You can specify actions that you want Sensitive Data Protection to take
when the inspection or hybrid job is complete. For example, you can configure a
job to save the findings to a BigQuery table or send a
Pub/Sub notification.

#### Inspection jobs

Sensitive Data Protection has built-in support for select
Google Cloud products. You can inspect a BigQuery table, a
Cloud Storage bucket or folder, and a Datastore kind. For more
information, see [Inspect Google Cloud storage and databases for sensitive
data](https://docs.cloud.google.com/sensitive-data-protection/docs/inspecting-storage).

#### Hybrid jobs

A hybrid job lets you scan payloads of data sent from any source, and
then store the inspection findings in Google Cloud. For more information,
see [Hybrid jobs and job triggers](https://docs.cloud.google.com/sensitive-data-protection/docs/concepts-hybrid-jobs.md.txt).

### Inspection through a `content.inspect` request

The `content.inspect` method of the DLP API lets you send data
directly to the DLP API for inspection. The response contains the
inspection findings. Use this approach if you require a synchronous operation or
if you don't want to store the findings in Google Cloud.

## Sensitive data de-identification

The de-identification service lets you obfuscate instances of sensitive data.
Various [transformation methods](https://docs.cloud.google.com/sensitive-data-protection/docs/transformations-reference.md.txt)
are available, including masking, redaction, bucketing, date shifting, and
tokenization.

There are two ways to perform de-identification:

- Create a de-identified copy of Cloud Storage data using an inspection job. For more information, see [De-identification of sensitive data in
  storage](https://docs.cloud.google.com/sensitive-data-protection/docs/concepts-deidentify-storage).
- Send a [`content.deidentify`](https://docs.cloud.google.com/sensitive-data-protection/docs/reference/rest/v2/projects.content/inspect.md.txt) request to the DLP API. For more information, see [De-identifying
  sensitive data](https://docs.cloud.google.com/sensitive-data-protection/docs/deidentify-sensitive-data).

## Risk analysis

The risk analysis service lets you analyze structured
BigQuery data to identify and visualize the risk that sensitive
information will be revealed (*re-identified*).

You can use risk analysis methods before de-identification to help
determine an effective de-identification strategy, or after de-identification to
monitor for any changes or outliers.

You perform risk analysis by creating a risk analysis job. For more information,
see [Re-identification risk analysis](https://docs.cloud.google.com/sensitive-data-protection/docs/concepts-risk-analysis.md.txt).

## Cloud Data Loss Prevention API

The Cloud Data Loss Prevention API lets you use the Sensitive Data Protection services
programmatically. Through the DLP API, you can inspect data from
inside and outside Google Cloud and build custom workloads on or off
cloud. For more information, see [Service method
types](https://docs.cloud.google.com/sensitive-data-protection/docs/concepts-method-types).

### Asynchronous operations

If you want to asynchronously inspect or analyze data at rest, you can use the
DLP API to create a
[`DlpJob`](https://docs.cloud.google.com/sensitive-data-protection/docs/reference/rest/v2/projects.dlpJobs.md.txt). Creating a
`DlpJob` is the equivalent of creating an inspection job, hybrid job, or risk
analysis job through the Google Cloud console. The results of a `DlpJob` are
stored in Google Cloud.

### Synchronous operations

If you want to inspect, de-identify, or re-identify data synchronously, use the
inline `content` methods of the DLP API. To de-identify data in
images, you can use the
[`image.redact`](https://docs.cloud.google.com/sensitive-data-protection/docs/reference/rest/v2/projects.image/redact.md.txt)
method. You send the data in an API request and the DLP API responds
with the inspection, de-identification, or re-identification results. The
results of `content` methods and the `image.redact` method aren't stored
in Google Cloud.

## Pricing

For information about costs associated with using Sensitive Data Protection,
see [Sensitive Data Protection pricing](https://docs.cloud.google.com/sensitive-data-protection/pricing).

## What's next

- Learn how to [profile data in a project](https://docs.cloud.google.com/sensitive-data-protection/docs/profile-project.md.txt).
- Learn how to [start or schedule an
  inspection](https://docs.cloud.google.com/sensitive-data-protection/docs/schedule-inspection-scan).
- Learn how to [inspect data from external sources using hybrid jobs](https://docs.cloud.google.com/sensitive-data-protection/docs/how-to-hybrid-jobs.md.txt).
- Learn how to [create a de-identified copy of data stored in Cloud Storage](https://docs.cloud.google.com/sensitive-data-protection/docs/deidentify-storage-console.md.txt).
- Learn how to [compute k-anonymity for a dataset](https://docs.cloud.google.com/sensitive-data-protection/docs/compute-k-anonymity.md.txt).
- Learn how to [de-identify and re-identify data using the DLP API](https://docs.cloud.google.com/sensitive-data-protection/docs/inspect-sensitive-text-de-identify.md.txt).