this post was submitted on 25 Apr 2026
33 points (100.0% liked)

Self-hosting

4324 readers
7 users here now

Hosting your own services. Preferably at home and on low-power or shared hardware.

Also check out:

founded 4 years ago
MODERATORS
poVoq@slrpnk.net
sam_uk@slrpnk.net
33
Hister - Your Own Search Engine (hister.org)
submitted 2 weeks ago by monica_b1998@lemmy.world to c/selfhosting@slrpnk.net
9 comments fedilink hide all child comments
top 9 comments
sorted by: hot top controversial new old
[–] Sibbo@sopuli.xyz 11 points 2 weeks ago* (last edited 2 weeks ago)

So did anybody try this and wake to share their experiences?

  • Does it use lots of CPU, RAM or disk?
  • Are the search results actually good?
  • Does it use a browser extension or so to get new visited sites, or do I have to import my history every day?
  • Does it also have a crawler?

Also, why would I use this over e.g. YaCy?

[–] eutampieri@feddit.it 5 points 2 weeks ago (1 children)

What’s the difference between this and SearXNG?

[–] uuj8za@piefed.social 3 points 2 weeks ago (2 children)

Huh, just from browsing the homepage, this seems to be more for searching local files.

[–] eutampieri@feddit.it 2 points 2 weeks ago

Hmm, thanks. I did browse the page but I didn’t understand well that aspect

[–] xnx@piefed.social 1 points 2 weeks ago

It’s not. It makes local versions of websites you visit so you can search them

[–] activistPnk@slrpnk.net 4 points 2 weeks ago (1 children)

I currently use the find-grep function in emacs, which is basically: find . -type f -exec grep 'my.*search.*pattern' {} +

To do PDFs, I use something like find . -type f -iname \*pdf -exec pdfgrep 'my.*search.*pattern' {} +

My problem is generally when TOKEN1TOKEN2 has a line break between tokens. It’s fucking annoying that grep is line-by-line. I wonder if Hister solves that problem. But from the website, I see no advanced syntax. I would love to search a pattern like word1 w/s word2, which would find cases where word1 and word2 appear in the same sentence. And word1 w/p word2 to match cases where two words are in the same paragraph.

[–] cravl@slrpnk.net 1 points 2 weeks ago (1 children)

Replacing line breaks with nulls first is an option. That's a lot of extra processing for very large blocks of text though.

Using regular grep is possible with the right flags, or you could also use pcre2grep with the -M flag, which should be available on every distro nowadays. See this Stack Overflow article for details.

[–] activistPnk@slrpnk.net 1 points 2 weeks ago* (last edited 2 weeks ago)

pcregrep is not automatically installed with Debian but it’s in the official repos. It seems common to get:

pcregrep: Too many errors - abandoned.
pcregrep: Error -8, -21 or -27 means that a resource limit was exceeded.
pcregrep: Check your regex for nested unlimited loops.

But it will help in many cases. I can see that it works on sufficiently small files. I noticed the built-in grep function for emacs can be modified to use pcregrep w/-M added instead of grep, which I find quite important because emacs makes it very easy to jump around to visit different results. In the end it’s still a hack.

[–] Taleya@aussie.zone 2 points 2 weeks ago

Ooof. I'll be honest. I head hister, i think about that nostradamus theory.