Meshes generated from reality, especially when one needs them to be really high quality, can be as huge as multibillion polygons, size that can be barely read in majority of CAD or modeling packages. Memento goes beyond that - it uses the same cloud service behind the scene to create the same high qiality 3D mesh, but then it has a toolset to clean up precisely the surrounding however needed, find visible and invisible errors on the mesh ( holes, spikes, particles, tunnels(, has tools to fix them automatically or manually (mesh diagnostic or the manual tools such as hole fill, bridge, ), can measure, scale and compare meshes, can do basic scultping where fix is needed and extrude to create a full volume when needed, calcultes surface, volume, has a 3D print toolset, has tools to retopologize and decimate the mesh and most importantly a variety of ways to prepare the clean fixed mesh for various downstream uses. If that is good enough for you and you dont need any further edits to the mesh, reCap would be good enough for you. The mesh creation is great quality, but there are no further edit fix or prep tools. You upload photos and get a high quality raw 3D mesh geometry generated by the photo to 3d cloud service that you can export as is (or just roughly cut by a cut plane) to different formats the automatically generated mesh. Photo on reCap360 is photo to 3d mesh service, full stop.

The difference between Photo and Memento today are the followng: In Autodesk, we have one photo-to-3D engine, and we use the same engine behnd different products (123D Catch which is consumer grade and limited in quality, Photo on reCap360 and Memento).

To add a bit more detail of what Mitko started explaining.

Autodesk memento how does it work