Spotlight doesn’t currently appear able to search Time Machine backups reliably, at least not in Sequoia or Tahoe, although this may not be universal.
[…]
If I now step back through my backups to reach one that I know contains that file, I can restore it. But if I type anything into the search field, nothing is found. If I change the scope of the search to that backup, the window title changes but its contents remain blank, and there isn’t even a busy spinner to indicate a search is in progress.
[…]
As I can’t disable Spotlight indexing on that volume without macOS telling me that it’s required to be indexed by Spotlight, neither can I force that volume to be reindexed.
[…]
My conclusion is that using Spotlight to search Time Machine backups no longer works, and the instructions given by Apple are also broken.
That has been my experience, too. As I see it, there are four separate problems:
I want to be able to search all my backup snapshots, because I don’t know which one might contain the file I want to restore. Time Machine only lets you search one at a time.
Searching an individual snapshot doesn’t actually work.
Spotlight indexing is invasive, using lots of CPU time, and with spinning hard drives causes wear and tear and creates lots of noise. If search doesn’t work anyway, I’d like to turn it off, but macOS won’t let me.
If the problem is with the index, there’s no way to reset it without deleting your backup.
Unfortunately, all those saved versions in the version database fall outside the scope of Spotlight indexing, and Spotlight search can’t look inside any of the old versions saved in a volume’s version database. Surprisingly, the version browser doesn’t offer any search facilities either, as that’s presumably another feature intended for a future that never came.
[…]
One way around this is to save each document version as a separate file, allow Spotlight to extract their contents and add those to its indexes for that volume, then to search those files. This is quick and simple using my free utility Versatility.
Apple today announced that it has rebuilt the search infrastructure that powers key features like Spotlight, Photos, and Mail across all of its major next-generation software platforms.
Level up basic search into a retrieval-augmented system using SpotlightSearchTool and LanguageModelSession. Explore Core Spotlight integration, delegate-based hydration patterns, and how metadata quality impacts your search results. Learn how to use custom PipelineStages for tasks like sentiment analysis. Discover best practices for indexing and building flexible, context-rich search experiences in your app.
So from what I understand from the WWDC keynote, Spotlight is actually going to work?? When I search for simple files, it fails the VAST majority of the time!
Semantic search is different, in that its matches aren’t as crisp and Boolean. Rather than working like a simple index, it’s more like a thesaurus in effect.
[…]
Rather than compiling more exhaustive sets of keywords, semantic search can broaden the scope to cope better. And because we can interact through Siri, we can fine-tune our search results by specifying the cattle should be black and white, perhaps, and combining conventional search criteria such as location.
To get this to work effectively, there are some limitations. Because semantics are so contextual and variable, this involves apps and Core Spotlight. That’s a big benefit to user privacy, as Core Spotlight’s indexes are separated by user and stored locally, although in places like ~/Library/Metadata rather than volume-based Spotlight indexes in the existing hidden .Spotlight-V100 folders. And unlike global Spotlight indexing and search, it requires apps to have code to support both tasks, as it can’t just happen by magic.
While I’m sure we’ll all be impressed with many of the results of semantic search, hits that we never expected to find, it’s going to prove harder to assess those that it misses.
Personally, in my experience, I’ve noticed that the new Spotlight is much faster than the one in MacOS 26 Tahoe. It opens with less delay, and finds the right app or file instantly when Spotlight in Tahoe seems to have had a slight delay.
Already that has revealed the qualities of what would be offered as a Top Hit in its ranked search, with scores such as
mrScore: 0.2516, mrScoreText: 0.6878, kMDItemDisplayName: 1.0000, mrScoreFreshness: 0.50000000, mrScoreResultType: 0.0700, freshnessDate: (null), pommesL1=0.0000, L2=-4999.0283
While most of those are opaque, it’s notable that ‘freshness’ is included in ranking. In contrast, searches made in the Finder’s Find window are unranked, and don’t calculate any of those scores, nor take into account freshness.
That show of respect for its customers may change with iOS/macOS 27. Reports suggest that, at least in the first beta, Apple Intelligence is mandatory[…]
[…]
“So what’s the problem?”, you might ask. Apple’s AI takes up several gigabytes of storage and leaves less headroom for RAM.
Those are small inconveniences, however, compared to my biggest gripe with Siri AI: It’s completely ruined Spotlight.
[…]
The new Siri-first interface that presumes that if you’re searching for anything but an app or file, you must want Siri to feed you a few links of Apple Intelligence’s choosing.
Getting to a web search from a Spotlight query now requires multiple taps: Type your query, tap “Show Results” (careful: hitting enter will trigger Siri to craft a response, eliminating the possibility of seeing any actual Spotlight content), tap on “Show More” next to the list of Siri-surfaced web results, scroll down until you see Search Google (or whatever engine you have set as your default), then tap that.
Apple Intelligence used to be opt-in, but now it seems that you can’t even opt out. I had previously mentioned the removal of the switch in the context of hypocrisy, but it seems there are enough issues here for it to warrant its own post. Also, I had previously written that the switch was combined with Siri, meaning that you could turn off Apple Intelligence if you turned off Siri, too. I now doubt that’s the case.
Depois de já termos publicado vários vídeos sobre as novidades do iOS 27, chegou a vez de começarmos a cobertura em vídeo do macOS Golden Gate 27 em nosso canal. E não tem jeito melhor de abrir essa série do que falando sobre a mudança mais aguardada do sistema: a Siri AI.
Neste primeiro vídeo, mostramos como funciona o novo app da Siri, os ajustes de voz disponíveis, as melhorias na busca do Spotlight, os recursos da Inteligência Visual a partir de screenshots e as ferramentas de escrita e revisão de texto integradas ao sistema.
I have a Mac Studio M2 Max with 32 GB of RAM and a 1 TB SSD in front of me. While not the most powerful personal computer you can purchase today, it is among the most computationally capable ever made—a vastly capable machine. So why do I have to wait for typing to catch up with me and see a rainbow spinner on the regular?
I’ll tell you why: iCloud Drive and Spotlight. Apple lacks any transparency about how iCloud Drive works or fails. The same is true of Spotlight. These big black boxes churn away, performing a million billion operations a day, and we don’t know exactly what they do. That’s not me being conspiratorial: I honestly don’t care what’s happening under the hood—until it breaks. Without diagnostic tools, even someone like me, ostensibly an expert, can be completely at a loss and waste many hours.
I wrote “Cloudy with a Chance of Insanity: Unsticking iCloud Drive” for TidBITS back in October 2023 about a multi-month odyssey with Apple technical support in trying to dislodge unsyncable files.1 That problem has barely surfaced since—an 80K file will sometimes get stuck for a while—and I thought Apple might have really solved syncing, at least as exhibited on my Mac. I receive dramatically fewer emails about iCloud Drive than I used to, which is another data point.
But Mr. Edge Case here can always find a way to break something.2 I’ve been working on a large manuscript for my upcoming book, Flong Time, No See, a revised and expanded collection of researched articles I’ve written about printing and type history over the last several years.3 I pasted in text that was typically copied from Web pages, and wound up with over 250 footnotes.
While putting together the files for this book, Apple released its Creator Studio system, and shipped Pages 15. After having no problems, I dutifully switched over. This was months ago, and as I continued to work on this reasonably large file—weighing it at about 60,000 words now—everything seemed fine.
A few weeks ago, however, I started to experience system slowdowns. Pages was taking up 12 GB of system memory! System load went through the roof! CPU consumption was outlandish! All signs pointed to corespotlightd, a long-time enemy of performance. That daemon handles background indexing for Spotlight, and the slightest thing wrong in a file or directory, or perhaps due to mild corruption in its underlying files, turns it into a rampaging beast that eats processor cycles like I consume potato chips.4
There’s a well-known procedure for killing this beast, though another is spawned from its remains: stop the index, delete the index files, restart, and re-enable indexing. From the Terminal, use these commands. Make a backup before proceeding. Take extreme care with rm -rf, as a misstep in entering or copying can wipe out files!
sudo mdutil -a -i off
rm -rf ~/Library/Metadata/CoreSpotlight
You can use Restart from the Apple Menu while holding down the Option key (to avoid being asked), or you can take a shortcut from Terminal, assuming you have ensured all files are saved:
sudo reboot
After your Mac restarts, go back to the Terminal and re-enable indexing:
sudo mdutil -a -i on
For many (perhaps most) people, this solves the problem. For me, it just delayed it. With some isolation using System Settings: Spotlight: Search Privacy, I added specific folders, such as mail archives, since that seemed like it might be part of the problem.
Search Privacy lets you exclude volumes and directories from Spotlight indexing, which can remove unwanted results and sometimes (only sometimes!) reduce thrashing.
One piece of advice I found suggested that if I thought Pages was the problem, I should exclude the iCloud Drive path containing Pages files. Fine! I clicked the plus icon in Search Privacy, pressed Command-Shift-G to bring up the path dialog, entered ~/Library/Mobile Documents/com~apple~Pages/, and clicked Choose. This seemed to work for a while, then nothing availed.
Finally, after analyzing logs and caches, the culprit was clear. It was not that large a Pages file: every time I saved, Spotlight’s worker bee was performing excessive re-indexing that consumed gigabytes and one or more processor cores. Moving the file to a local directory (in my Documents folder) while editing eliminated the problem. It also freed up space, as the temporary Spotlight indexes and Time Machine snapshots had become absurdly large.
289 footnotes seem to be fine for Pages, but Spotlight and iCloud Drive find them hard to swallow.
What could Apple have done to help here? This is where I would hope machine learning, coupled with some interactive code- and system-expertise bots, could help:
Note that a system component is acting well outside known parameters.
Identify CPU usage that is tied to Apple-owned processes.
Provide a troubleshooting tool, even suggesting it in notifications.
Create human-readable logs that don’t require system administrator knowledge to interpret.
In the case I describe above, an Apple-created daemon was rampaging out of control due to a single identifiable file!—cascading into other problems. It should be solvable by Apple. The fact that we have more power under the hood by orders of magnitude in the past, coupled with neural cores in the CPU packages, just makes this even more embarrassing. Solve Siri, Apple, yes, but real transparency and troubleshooting in macOS would also make a big difference.
[Got a question for the column? You can email glenn@sixcolors.com or use/glennin our subscriber-only Discord community.]
I should have made an unsinkable/unsyncable pun in the title, but there is an esprit d’escalier moment for everything. ↩
An “edge case” is something that is often considered unlikely to occur with a product, whether software, hardware, or mechanical. Yet, no matter how much I’m just trying to get something done, I apparently trigger edge cases all the time. Software companies “love” me for this. ↩
Flong is a kind of paper-like mold that was extremely useful during the days of metal-based relief printing. Of course, you can pre-order a copy to get the full scoop! ↩
Login
