People, places, and trips

Avlok organizes your library into the places you went, the trips you took, the people you have named, and the things that show up. These tools let an assistant browse along those lines.

get_photos_by_location

Find photos taken near a place.

Inputs

Field	Type	Required	Notes
latitude	number	yes
longitude	number	yes
radius_km	number	no	How wide to search around the point
limit	integer	no	How many to return

Returns photos taken within the given distance of the point, in the standard shape. Photos without location information are simply not included.

get_locations

The places you have photographed, worked out from photo locations, grouped into recognizable cities and regions.

Inputs: none required. An optional country filter narrows the list to one country: a two-letter code like US works on both surfaces, and the cloud surface also accepts the country’s English name.

Returns a list of places with how many photos you have from each. An assistant can use this to suggest “you have a lot of photos from Lisbon, want to see those?”

get_trips

The trips Avlok has detected, based on when and where photos were taken away from home.

Inputs: none required.

Returns a list of trips, each with its dates and place. An assistant can then pull the photos from a given trip.

get_people

The people you have named in Avlok.

On your terms

Naming people happens on your device, and the face data Avlok uses to group people never leaves it. On the cloud connection, only the names you assign (and which shared photos each named person appears in) sync so this tool can answer. It surfaces only the names you have chosen to assign. Depending on your settings and your region, this feature may be limited or turned off. It is never on without your action.

Inputs: none required.

Returns the list of names you have assigned, so an assistant can answer “show me photos with Sarah.”

get_entity_clusters

The recurring subjects and things in your library, grouped together. For example, photos of dogs are grouped under one heading even if you would describe them in different words.

Inputs: none required.

Returns the list of subjects with how often each appears.

get_connections

What tends to appear together in your library: people who show up in the same photos, places that go with certain subjects, and so on.

Inputs: none required.

Returns the strongest of these pairings. This powers the kind of insight an assistant can offer about your library as a whole, rather than a single photo.