Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is awesome!

One of the features I really liked about MacOS is spotlight. Is there some (fast) equivalent that I’ve been missing on Linux? I’m aware of `locate`, but that only matches the names of files, not their content. Is there a search engine that indexes local content as well?



I believe Baloo (a KDE project) does exactly that (indexing local content). There are different tools you can then use to search for indexed data: I assume KRunner will be the closest to Spotlight, but you can search for files right from Dolphin, if you use it.

To be honest, I've found Baloo to be a resource hog in the past, so I've now the habit of disabling it right after installing a distro which comes with it (eg. Kubuntu) or not installing at all otherwise (eg. on Debian or Arch). I should probably give it a second chance, though: on paper, that's the right approach to search.


Baloo ist basically Spotlight and probably more but dare you are not that typical happy smile buisness user - have a few kernel source tree's laying around in your /home and you'll likely are getting friends with Baloo sooner (spinning rust) or later (ssd) because that beast manages to saturate even SSD IOPs and while doing so hands out segfaults on every other file.

You can switch to disable content-indexing and just search for filenames (80% rule) but it's hidden in systemsettings or you even need to go to balooctl - also index corruption is a thing...

That beeing said, if you are a brave KDE user and don't mind the hassle it might even work 80% of the time - in theory it's great idea but it never worked reliable for me - if you are a bored dev there is probably lot's of low hanging fruit there - seccomb, some simple heuristic to not overload on IOPs, there are probably more efficient db structures than LevelDB and so on...

It's step forward from nepomuk? that did the full rdf semenatic web stuff and fed a relational database and killed your hdd-based desktop reliable in the early 2000ies but it's still a nasty surprise when using KDE.

It's still a cool idea but it needs some love and contributors to work reliable on all kinds of nasty setups.


I’ve found searching for files on KDE (with Baloo) to be, for some reason, really inaccurate and hit-or-miss. Sometimes, even if I remember the exact name of the file, it won’t show up in the search results. Partial names have a lower probability of showing up. Slightly inaccurate names (with a Levenshtein distance of 1 or 2 to the actual name) have a very low probability of showing up. I end up searching for it using find or ag (the silver searcher).

Baloo is also indeed a resource hog. It can use 100% of a CPU core for hours and hours, while it’s indexing. But I’d be fine/happy with that if it just worked properly.


Ah, I can confirm all of the issues. I've thought that was my fault - so these are probably bugs.


Baloo is alphabetically first dependency of Dolphin in Debian (both Stretch and Buster), so either you have to use a different file manager or disable it after install.


https://wiki.gnome.org/Projects/Tracker

Most gnome based distros have that available, if not enabled by default. Not sure what kind of support it has for different file types.


There's Recoll: https://www.lesbonscomptes.com/recoll/

I haven't used it, so I'm not sure how fast it is, but I've seen it recommended several times.


Recoll is excellent. It was a game changer for me once I finally set it up on a NAS at home with a web interface accesible by port forwarding over ssh. Though the web interface requires some tiny tweaks to be mobile responsive.

+ It searches within compressed file types recursively.

+ Searches damn near everything

+ The huge number of way to interact with it - gui, python, command-line, web interface - combined with extensive if kinda weird query language make it clear its been refined for a long time.

+ Windows GUI

- Pain in the ass to make work right on windows, and the indexing on windows seems to be way longer for some reason.


Recoll is great. I use it like gmail search for my computer-global file system and the files that get put there...who needs extensive folder organization when you can find anything, anywhere?


Only for pdfs (and I use/develop it specifically for academic pdf): https://github.com/bellecp/fast-p





Consider applying for YC's Fall 2026 batch! Applications are open till July 27.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: