Grimoire: Open-Source bookmark manager with extra features
github.com
github.com
The devs are also ex-Pocket users and have worked hard to get feature parity and then some. There are mobile apps too for reading on the go (and work offline) which I use extensively when I am on flights. There is a graphql API and webhooks you can use for extending its functionality. Search could be a little better, but I use the labeling system which works well. I also use the logseq integration to keep a persistent log of articles I read on any given day.
Thanks for any input
As per https://grimoire.pro/docs/roadmap sharing, import, export, bookmarklet and browser extension are on the roadmap.
Offtopic: I've heard the word Grimoire only from the Black Clover anime and expected the author to be a fan of the series. Didn't know that it's such a common word, so TIL.
And, local -- thank goodness. Web apps weren't really even a thing, yet.
This also meant a compact, information-rich GUI, as opposed to all the low-density web crap we're forced to put up with, these days.
It's EOL. IIRC, in part browsers closing themselves off to third party application access did them in.
Anyway, if anyone's serious about making a good personal bookmark manager, take a look at PowerMarks, if you can somehow. No one else I've seen has come close to what they had.
I'm very flattered by such good observations and the warm welcome for this project. I won't be able to respond to every comment from you guys, but you bet I will read them all and take notes!
For now, Grimoire is at 1/10 of its potential, and I will work hard to make it worthy of being a good contender for your default bookmark manager. Definitely, it's missing many features, like a dedicated browser extension, import/export capabilities, and better documentation, and I'm well aware of that.
Was the launch rushed? Maybe, but I thought it would be great to hear your opinions and perhaps even appeal to some potential contributors (wink wink).
For now, I want to address the most common issues that prevent some of you from even running and testing it (clearly an oversight on my side). Then I will write a blog post on https://grimoire.pro to answer some of your questions and doubts, so (if not now, but maybe in the not-too-distant future), give it a chance.
Thank you again, and big kudos to user hunderbong for mentioning Grimoire on HN!
The UI browsers provide for bookmarks is pretty atrocious, and seems to have been the main reason I didn't go back to use them much - which I didn't realize until I had the idea to use Zotero as a bookmark manager. It has a lot of the features that Grimoire lists, including separate "user accounts" - I have a separate Work profile, for which I've created a desktop shortcut the same I do for Firefox [1], with its own collections and tags.
I still welcome any new well-built alternatives like Grimoire, and I agree with you that a first-class UI is important.
I don't have a use for Grimoire - not willing to change my bookmark workflow at this point - but you can be sure I'll remember it just based on that.
Who enjoys opening their bookmarks and seeing 500 links in an unordered list? Metadata has been a thing for how long now, and yet no browser that I know of has gone ahead and implemented a proper bookmark management workflow.
But what you really need is a google-like, or even better, LLM-powered search for the locally cached text of the pages you bookmark. README for this software mentions fuzzy search, not sure if it's capable enough.
That's exactly it for me. Once I get a few bookmarks in the list, I'll inevitably need to organize them, but then I'll find I don't want to take the time to do that, so I just don't bother saving the bookmark.
At the end of the day, I just want to dump all my links somewhere and be able to find them again. I don't want to think about where they should go.
I built a journaling app recently [1] (the UI is similar to twitter, but for your personal thoughts), and what I've ended up doing a lot is dumping links in there. What I've found that helps is writing a quick note and adding tags, but even more than that, having a link preview generated makes it easier to remember what the link was later. (Based on screenshots, it looks like Grimoire is also generating a link preview.)
The fact that the journal UI is a "feed" also means I can browse through my history and see when I saved the link, so it has a chronological aspect to it as well. I think having all these affordances for finding things (commentary text, tags, link preview, chronology) is way more useful than hierarchical management.
You might want to check out xBrowserSync. Its open source, can be self-hosted, has browser extensions and an Android app.
You don't have to organize anything. When you add bookmarks, you add them with tags and they will all go into the default bookmark folder. The beauty is that you never have to interact with that folder. Instead you use the browser extension to search for tags, name, and description. Its a very simple app that does 1 thing well.
The UI does look a little dated, but there's so little of it that it doesn't matter much.
That is why I included search on my bookmark manager. I use it daily. It contains search.
It fails to search anything, even it's own folders! or parts of the url. or the tags that gets added and removed from the UI every now and then and are ignored in search.
heck there's a 18yr old bug about performance being so bad in this search that if you have over 1k bookmarks it will just hang for several seconds.
edit: that said, i also use firefox bookmarks exclusively, but just for the (only one true in the entire industry) e2e server backups. Then i just manually export them and search outside of firefox, which sucks.
That said, all you have is the url and description if you don't tag, so yeah, it's hard for it to be really smart about what you are looking for. As another commenter mentioned, this is an area that an AI assist tagger or search would be useful. But the industry want to drive you to a search engine so you can be data mined and shown ads.
* If I have over 1k bookmarks, that probably means it's time to go in for spring cleaning and start deleting some
* Other comments have made it apparent that you can tag bookmarks, but I personally just prepend the names of bookmarks like "javascript js - How to reverse a singly linked list GeekForGeeks" to make it easier for me to search for it in the future.
In the end, I prefer the benefit of not having an extra site/tool like Instapaper/Pocket/Shiori to have to deal with.
Want to open Facebook? Search 'Facebook' and click on the link.
Want to open your bank site ? Again google bank name and click on the link.
I've see even tech people do this. The only bookmarks I've seen my colleagues use are for internal reference wikis and links which are not readily searchable.
I don't know whether this is a search or UX or people problem but the whole workflow can be improved for sure.
It's UX, but Google and Bing want it this way, so...
[0] https://en.wiktionary.org/wiki/<term> and https://docs.rs/<lib> come to mind.
If you are doing this often enough, may be adding a custom keyword search might help. Both chrome and Firefox support it. Very useful with random JIRA and other ticket/id type navigation to skip search and jump directly to their pages.
> maybe that's subconsciously associated with my impression that Google Search results have gotten worse and worse over the years.
It has gotten worse and I don't think even Google will contest that claim.
But they will search less. Which, as we have seen in the recent Google anti-trust case, is also a point Google looks at intensively when they introduce new features.
it almost work like I can access all things I need by "search" with Google...until I found Google does not remember anything after two year.
Docs (folder, expanded)
Node
FS
Path
SQLite
Database
Statement
Syntax
…
Metadata has been a thing for how long now
It would be naive to assume that it will contain a systematized coherent description of a page content.
Now there are bookmark tools which are archiving a page for example, have full text search. And you can add more and more features, it’s an endless game - that’s why we have that many bookmarking tools. Everybody has different priorities and needs different things.
Other browsers copied their homework. Makes less sense.
Just wanted to congratulate you on this, looks great! Especially the bookmark cards and user profile page / infos. Will probably be copying some of that down the line :wink:
What would be awesome is if I could wget all the assets locally, and then use a local LLM to search its contents.
Stores using it because I use several computers and two OS now, and the utility of it is pretty low for me.
You could make a donation to Archive.org and just a script to check for an "archive-this" tag in your bookmarks, then pass those pages to archive.org??
(It does have some idiosyncrasies since it's supposed to be a citation manager originally, but once you get used to the workflow it gets out of the way and works really well.)
Interesting to see this project pull it through anyway.
They also begin with updating a lot of deb-dependencies. I imagine this to be hellishly difficult to maintain. Props for even trying!
I worked on something in 2007 for bookmarking, too. https://vimeo.com/60251128
I use self hosted solutions whenever possibile. For bookmarks I wrote a special program for myself for that, as no other program, service fully ticks all the right requirements.
Docker is not without complications, but it alleviates far more problems than it causes. I get annoyed when something is released and doesn't have an official Dockerized version. Especially when it's a web service.
I'm a Python/Ruby/Javascript developer, and I don't want any part of those installed except on my dev machine. Unfortunately Python has infested too much of Ubuntu to avoid.
What's really nice is using direnv to cd into a directory and having a flake.nix file in it which changes the entire system to what I need to do in that folder. So I don't need any of my dev environment to be system-wide, everything is per project. That alone has cut my system down to pretty much just system tools (replacements for shell commands), productivity tools (Obsidian, bitwarden, etc), and my browser. It's just so nice.
How cross-platform is direnv + Nix? Do flake.nix files have macOS/Linux versions?
One of the biggest advantages of using Docker images as a package format is that it's (sort of) cross-platform via Docker desktop and such on a Mac. I would prefer to use something like Nix if possible if it solves the problem a different way (being aware of platforms).
Beats the tar out of the obnoxious curlpipe pattern anyway.
- People who use Docker (Compose) can easily get up and running with a project in minutes
- People who dislike/don’t use Docker can at least use the Dockerfile and Docker Compose file as a reference to learn how to install the dependencies and the service itself.
I kinda prefer docker installations because I don't need to guess what implicit "works on my machine" dependencies someone had, and the whole thing can be mostly stateless or the extra state is made pretty obvious.
For example, a study on this was recently conducted for PyPI. https://arxiv.org/abs/2309.11021
An updated Docker environment is pretty secure I think.
Right now I use xbrowsersync. It's got major problems, like sometimes (always?) breaking the ability to add bookmarks at all when sync isn't working, having to authenticate with a long-ass sync ID that you can't possibly remember, and not remembering recently-used bookmark folder order. But it works on the browsers I use and can be self-hosted. I'd love an alternative with a better UI that fully integrates with the browser (in my case ungoogled Chromium).
A big plus is that is a single Golang binary so very easy to set up. It also has a browser extension.
That mens I cannot periodically import my Raindrop bookmarks to shiori. After I've done the import I can either add new bookmarks to both immediately and by hand. Or I do a total re-import in shiori, which re-archives everything (and if a page isn't retrievable anymore, that's bad luck).
It appears to be impossible to have a "public" page? How it does everything else is perfect for my style, but I'd like to be able to share my link page to my students without login?
Just save the complete page, only selected text or only the link to a markdown file or Obsidian. With downloaded, linked or without pictures. My OS and Obsidian can search those files, they have more (automatically added) metadata.
I can even edit them in the browser: add your thoughts, tags or change the name of the file before they are saved.
I can (automatically) do with them what ever I need. They can be used to (automatically) generate an always up to date start page or a data vault on GitHub.
My local AI assistant can parse them.
Local, versatile, permanent, flexible, cost effective, future save. No need for a bookmark manager.
Could you elaborate more on the "local AI assistant parsing" part please?
This has made my digital life a lot less burdensome to manage, and it also felt freeing to drop all that unorganised baggage.
To all those bookmark junkies out there - do you really actually read all that stuff, or is it just data hoarding? (And I don’t mean this in spite, I’ve hoarded lots of data at one point or another myself :-))
Since I never know what I will actually find useful at a later time, I spend no energy organizing them. But having a simple record of my (want to) reads is super useful whenever my brain vaguely remember having seen or read something that comes up in a conversion/project, but not where and the exact name. Since I use broad on-the-top-of-my-head tags, I usually find what I'm looking for within a few seconds.
This strategy has lowered my barrier to bookmark something significantly, and reduced the mental overhead to almost zero.
Decades of overthinking my bookmarking habits and this has been by far the most useful system!
I was a big del.icio.us user back in the day, maybe self hosting something like the OP is the way to go.
...
Isn't submitting a response enough to convince the rest of us plebs that you agree or have other strong feelings about the parent comment?
Chances are good that, unless you're James Joyce, folks will read through the entirety of your statement to draw inflection from what you've conveyed.
When you don't agree with the parent comment, have you kicked things off with, "Not this" or, more succinctly, "!This" before making your counterpoint?
For me, Google has a mind of its own and has not been reliable in giving out related links even when I search with the exact page title from memory. Happened a lot with stackoverflow type sites.
I use Firefox bookmarks and think it has a reasonably good bookmark manager with support for tags and custom keyword searches etc.
Tip: In Firefox you can search bookmarks directly by prefixing searches with a * . It can even search bookmark tags.
Sadly Edge doesn't do tags, I initiate them by appending keywords to the bookmark name.
(I have to use Edge at work.)
Additionally I have this page https://david.kitchen/bookmarks/ which is my browser home page in all browsers, and is another form of a bookmark collection, and this is how I mostly interact... where pinboard is long-term storage of bookmarks, anything I interact with regularly enough is on this page. I update this whenever I find myself returning to the same thing for more than a few weeks.
I don't feel like I carry any baggage, the opposite in fact... I feel like this clears the mind of the responsibility to remember as this is mostly long-term memory and short-term convenience.
I also keep my browsers in permanent incognito mode with NoScript... so tab-city is not a thing I've ever done, every session is a blank session. I view bookmarks like data on a computer... it should be extremely portable across devices and OSs, and so bookmarks should transcend the device, a specific browser, and be available everywhere.
"Firefox is the Chrome replacement I've already been using."
I then keep one directory for stuff that is on my short term future "todo" list. For example, when I am deciding on buying a product I might have multiple shop listings there.
So to your statement here
> To all those bookmark junkies out there
My response is that there is a middle ground ;)
First off, I almost never encounter dead links. That’s an absolute rarity and I can’t even remember when that last happened. That’s more for reddit/SO links.
Some that I have as favicon only for regularly visited sites, but not regularly enough to have them as pinned tab.
Following those, I have temporary bookmarks. That’s currently fantastic fiction as a reminder to myself to use it, that will probably be gone next year as I got used to using it.
Then, folders. First is cooking stuff, from before I used a recipe manager. That’s around 100 bookmarks there that are categorized into subfolders.
Music. Mostly YT vids that are not on Bandcamp where I have a wishlist. Usually use those for songrequests on twitch.
Next is the voice folder, a long-term project of replacing alexa with something fully local, everything I researched or even just hardware sources.
Sesotho, for information for learning the language. Only relevant when I’m actually studying something new there.
3 gaming folders for build inspiration from tabletop guides (WotR, Solasta, BG3; videogame guides are way too minmaxy for me)
Following those are random links. I’m now roughly at 1/3rd to 1/2 of my window size, everything that follows is currently relevant stuff that I still want to lookup or make use of. New ones get added to the left. Once something falls into the overflow area, it’s probably irrelevant, but often I do get back to those links before that happens.
So yeah, I love bookmarks.
But yes, over two and a half decades of storing URLs with little context other than their place in a hierarchy of topics has led me to the conclusion that data on the internet is ephemeral. If you want to save it, copy it.
Even stuff I've read, if I remotely think I'll want to see it again, I bookmark and tag.
I never manually go through my bookmarks, only search them. But they are a pointer back in history for me.
Off-topic, but given how people try to apply LLMs to everything right now, I am surprised there is no extension or something similar to help manage bookmarked and open tabs.
Anything else is purged.