It's 2026. If You Still Miss Picasa, Try Picacino.
Google shut Picasa down on March 15, 2016, and the photos stayed behind in ordinary folders. Picacino reads those folders the way Picasa did, offline and without an import, and writes your corrections into the photos themselves.
You probably have years of digital images sitting in ordinary folders. Picasa was a desktop photo manager from Google that made sense of that exact structure, scanning the directories already on your drive and building its own database while tracking edits in a hidden .picasa.ini file. It had face recognition with name tags, and it worked, right up until the day the desktop app was discontinued on March 15, 2016, so Google could focus on a single service called Google Photos.
The folders exactly as you built them
Picacino is built around that original idea, reading the folder tree you already have without requiring an import step or making you commit to a catalog. Nothing is copied, moved or uploaded. The sidebar shows your exact structure with a photo count on every branch, and because the folders are watched, dropping a new photo into one makes it appear on its own. The folder on disk remains the source of truth. Picacino's database is only a cache you can delete and rebuild without losing anything but the time it takes to read the files again.
Reading the library happens in the background, so a hundred thousand incoming photos will never freeze the window you're scrolling. The grid only draws what is on screen, generating the thumbnails you are looking at right now first, so browsing works within seconds of adding a directory. Right-click an image anywhere in your file browser and choose Open With Picacino. The folder it sits in joins the library, the sidebar opens to it, and the picture comes up full screen as a slideshow through the rest of that directory whether the app was already running or not. It reads JPEG, PNG, WebP, GIF, TIFF, BMP and HEIC.
Looking through a shoot
You get two ways to look at a folder: a grid with a size slider, and a list with sortable columns for name, date, dimensions, size and rating. A detail panel on the right describes whatever you picked in either view. Opening a photo full size brings up a filmstrip underneath to help you move around a shoot, where you can scroll to zoom and drag to pan. Press 1 to 5 to rate, 0 to fit the image back to the frame, arrow keys to step, and Esc to go back. Any view plays as a slideshow.
Stars are the simple "I like this" mark, keeping you from having to maintain a second favourites list in step with your actual library. A handful of useful tags is there on the first run, so you delete the ones you don't want and add your own. Search is full-text over file names, cameras, tags and folders, letting you type a word or narrow it with prefixes like camera:, tag:, folder: and file:.
Recognition on your own machine
The People view groups photos by who is in them, giving you space to name a group, merge two groups that turned out to be one person, and take a face back off someone it filed wrongly. It reads every photo once, entirely on your own machine, using models that ship inside the app and take up just 9.5 MB. It only starts when you ask it to, and if you switch it off in Settings, it leaves entirely. On the machine it was measured on, it processed about 55 photos a second at 12.6 megapixels, which clears 10,000 photos in roughly three minutes. The scan runs once per photo rather than every time you open the app.
Corrections written to the file
Turn or flip a photo in the detail panel, check it, and hit Save to write the orientation tag the image already has room for, so the picture is upright in every other program and anywhere you send it without ever re-encoding the pixels. Correct the capture date for a camera whose clock was set wrong, alongside the camera make, lens, exposure, location, author, copyright and description. This is written into the photo file itself through a staged copy that is then swapped in, meaning an interruption leaves the original exactly as it was. The metadata editor works on one image at a time on purpose, so a single action never stamps the same values onto a hundred different photos.
Duplicates are found by content rather than by name, catching the exact same photo saved twice under different titles. Near-duplicates are caught too, finding a resize, a re-export, a second JPEG compression, or two frames of a burst even though not one byte matches. You get three sensitivities ranging from exact re-saves only to anything that looks alike, and every set says how alike it thinks the photos are, showing them side by side so you can choose which one stays.
Removing a photo takes it out of the library and leaves the file on disk. Put it back whenever you like. Deleting for good is a separate, deliberate step. Drag photos from the grid onto any folder in the sidebar to move the real files, taking their stars, tags and named faces with them. Anything that cannot move because a name is already taken in the destination is reported by name and left where it was. The notice has an Undo. Ratings, tags and which photos are in the trash are the things a rescan cannot rebuild, so Picacino saves them to a single file you can back up, carry to a new machine, or restore from.
Left out on purpose
Picacino does not crop, retouch or adjust colour. The detail panel has an Edit row that opens Pixpresso, and a Compress row for TinyImage. It runs offline: no web albums and no cloud sync. Picasa's .picasa.ini files and face name tags are ignored, so old edits do not come across. The photos themselves are ordinary files. Picacino reads those.
Writing metadata only works on JPEG, PNG and WebP. HEIC is left out on purpose: the underlying writer collapsed a tiled HEIC image down to a single tile. Rotation skips GIF, BMP and TIFF, which have no orientation tag. Faces are not available on Intel Macs. On Windows, HEIC requires Microsoft's free HEIF Image Extensions.
Whenever you're ready to look through those folders again, you'll find it on the Americano shelf, included in the free trial.