We’ve added a new section to the API documentation to publicly clarify our content safety guidelines:

By default, endpoints set the content_filter to low, which guarantees that no content violating our submission guidelines (like images containing nudity or violence) will be returned in results.

To give you flexibility in filtering content further, set the content_filter to high (on endpoints that support it) to further remove content that may be unsuitable for younger audiences. Note that we can’t guarantee that all potentially unsuitable content is removed.

We’ve added support for content filtering to GET /search/photos and GET /random.

param Description
content_filter Limit results by content safety. (Optional; default: low). Valid values are low and high.

Note: if you were previously using the experimental and undocumented safe boolean param, this has been deprecated and replaced by the content_filter param. safe=true corresponds to content_filter=high and safe=false corresponds to content_filter=low. We will be removing support for safe in a future update.