The HoudahSpot alternative for files that are pictures
HoudahSpot is the best set of controls anyone has put on top of Apple's search index. That sentence contains the catch, and the catch is the only reason to read further.
HoudahSpot is a precision front-end to the Spotlight index, so it can search anything Spotlight has already read, with far better controls than Spotlight gives you. Loupe builds its own index and opens the files itself, which is how a scanned page or a photograph of a receipt becomes searchable at all. If your documents are already text, HoudahSpot is the sharper instrument. If they are pictures of paper, it has nothing to point at.
What HoudahSpot actually is
It is worth describing properly, because most comparison pages do not. HoudahSpot takes the search Apple built into macOS and gives you the controls Apple left out: boolean operators, regular expressions, criteria stacked on top of one another, folders included and excluded individually, result columns you choose, and the matched text highlighted in a preview beside the list. You can save all of that as a template and run it again in six months without rebuilding it.
If your work involves the same awkward question repeatedly — everything over ten megabytes modified in March, by one author, anywhere except one folder — there is nothing here that competes with it, and this page is not going to pretend otherwise.
The ceiling it inherits
Houdah's own documentation is direct about the arrangement: Spotlight indexing has to be enabled for the app to work. HoudahSpot does not read your files. It asks Apple's index what is in them, and then does something far more interesting with the answer than Spotlight's own window does.
Which means the two apps share one limit exactly. Whatever Spotlight has not read, HoudahSpot cannot search, however the query is phrased. And the thing Spotlight has not read is the thing most people are stuck on.
A scan is a picture, not a document
When you scan a contract or photograph a receipt, what lands on the disk is an image inside a PDF wrapper. Open it in Preview and you can read it perfectly, because you can read pictures. An index cannot. Apple's image analysis does pull text out of ordinary image files, and has for a few releases now — but it does not do the same for the images inside a PDF. So a scanned page carries no text anywhere in the system, and a search tool built on that system is querying an index in which your invoice does not appear.
That specific gap has its own page, because it is the single most common reason people go looking for something other than Spotlight in the first place.
Words, not meaning
The second limit is quieter. Both tools match the words you typed. Search for the lease for the flat when the document is headed Tenancy Agreement and neither returns it, no matter how precisely the rest of the query is built. You did not misremember the file. You remembered what it was and searched for that, which is how people remember things.
What Loupe does differently
It opens your files and reads them once, in the background, and keeps its own index in a single file on your Mac. Text documents are read straight through. Scanned pages go through character recognition, so the words printed on them become words in the index. Pictures with no text are handed to a vision model that writes a plain description, and the description is indexed as ordinary text. On top of the keyword index it keeps a second one built on meaning, so a query and a document can match because they are about the same thing rather than because they share a word.
All of that happens on the machine. There is no server at the other end, which is why it keeps working with the Wi-Fi off. And it never writes: no file is moved, renamed, tidied or deleted, and there is no code path in the app that could.
| What you're trying to do | HoudahSpot | Loupe |
|---|---|---|
| Build a precise query: size, date range, author, folder, all at once | Its whole reason to exist. Attributes, boolean operators and regular expressions. | A handful of filters. Nothing approaching this. |
| Save a complicated search and run it again next month | Saved templates, which is what makes a recurring job a recurring job. | No saved searches. You retype it. |
| Find a Word document by a phrase inside it | Works, with the hit highlighted in a preview beside the result. | Works, and also finds documents that mean the phrase without containing it. |
| Find a scanned invoice by the supplier’s name | Nothing to find. Spotlight never read the page, so there is nothing indexed. | Reads the page with character recognition on your Mac, then searches what it read. |
| Find the photo of a whiteboard from a workshop | Finds words written on it, via Spotlight’s own image analysis, and tags it. | Describes the scene — three columns, arrows between them — and searches that. |
| Find “the lease” when the document says “tenancy agreement” | No match. Different words are a different query. | Matches meaning as well as words, so the document still comes up. |
| Run it on an Intel Mac, or an older macOS | Both, and has for years. | Neither. Apple silicon and macOS 26 or later, full stop. |
| Trust it with a decade of your filing | Over a decade in the field, still being maintained. | New. You have seven days to decide whether that bothers you. |
Where HoudahSpot is simply the better buy
Three cases, and they are not edge cases.
You are on an Intel Mac, or an older macOS. HoudahSpot supports both, going back many releases. Loupe needs Apple silicon and macOS 26 or later, because the models that read your pages run on the GPU. With Apple ending Intel support in macOS 27, the tools that keep serving those machines are doing something genuinely useful, and this is not one of them.
Your searches are structured, not vague. Meaning-based retrieval is good at the thing about the boiler and bad at every PDF over 10MB from last March. Those are opposite problems and it is fine for them to have opposite tools.
You want something with a decade behind it. That is a reasonable thing to want from software you are going to rely on, and it is an argument this product cannot answer for another few years.
The honest summary
They are not really the same category. HoudahSpot makes an existing index answerable in ways Apple never exposed. Loupe makes files searchable that no index on your Mac has read. If your filing is text and your questions are precise, buy theirs. If the things you cannot find are photographs of paper, precision is not what is missing.
Questions people ask about this
Does HoudahSpot use the Spotlight index?
Yes, and Houdah say so plainly in their own documentation: Spotlight indexing has to be enabled for HoudahSpot to work. That is not a criticism. It is how the app can offer hundreds of searchable attributes on day one without asking you to wait for an index of its own. It does mean that whatever Spotlight has not read, HoudahSpot cannot search.
So can HoudahSpot find text in a scanned PDF?
No, because there is no text in one to find. A scan is a photograph of a page, and Spotlight does not run character recognition over the images inside PDFs, so the words on that page were never indexed by anything. No amount of query precision reaches text that was never read.
Which one should I buy?
If you write the same complicated query repeatedly — everything over 10MB modified last March, outside Dropbox, by one author — buy HoudahSpot, and it is not close. If the files you cannot find are pictures of paper, that precision has nothing to point at, and this is the one that reads them. They do different jobs and plenty of people own both.
I have an Intel Mac. Can I use Loupe?
No. Loupe needs Apple silicon and macOS 26 or later, because the models that read your pages run on the GPU. HoudahSpot goes back many macOS versions and runs on Intel, which on an older machine makes it the only one of the two you can actually install.
Can I keep using both?
Yes, and they do not interfere. Loupe keeps its own index in a single file of its own and never touches the Spotlight one, so HoudahSpot carries on working exactly as it did.
Try it on the folder of scans.
Point it at wherever the scanned paperwork ended up and search for a supplier you only half remember. Seven days, no card, and nothing of yours leaves the machine.
- Apple silicon · macOS 26 or later
- No account
- Nothing leaves your Mac