I like Raindrop. Developed by an indie dev, no ads and tracking, revenue comes from subscriptions.
It has a strong privacy policy, full API access to all your data, and the ability download your data as a zip file. When you delete something / delete your account, it is truly gone, not soft-deleted.
It is really fast (no bloated SPAs) too.
It can automatically import your bookmarks, and will make ongoing knowledge management, sharing etc seamless.
The extensions are open source: https://histre.com/install/
As a demo of how histre automatically organizes your knowledge, check this out: https://histre.com/hn/ (Filter HN by automatic tags)
(Disclaimer: Founder)
We’re solving knowledge management and collaboration in teams. At an individual level, bookmarks are an important input for knowledge. So we solve that.
> “Supercharing teams with AI” What does that even mean
We use gpt4 to summarize and auto organize knowledge: which could be from team chats, documentation, bookmarks, etc. This helps teams collaborate better without having to play Sherlock on team chat logs etc… hope that makes sense…
I do agree with you that I need to do a better job of describing it. I’ll work on it. Thanks for your feedback!
My understanding is that anything you hit an OpenAI API with can and will be used as training material for future public models. Thus it’s not safe to use their APIs with any proprietary data.
Your service seems pretty cool, but I’m nervous about connecting a pipe from my corporate chat solution to OpenAI like that.
No user data is automatically sent to gpt4. It’s strictly opt-in. Chat is sent to gpt4 only when the user specifically requests it.
Encrypt the file if you want extra security and use a good password/password manager.
I understand it's not elegant UX/UI though.
Edit: chrome offers to export to html - might be a better start! You can then turn that into txt or copy paste what you're interested in!
<a href="link">title</a>
Or even better: could just make a little self hosted web app that does this for you
Or another idea (most useful for me): self host an instance of mediawiki and use it as a personal knowledge base
[1] https://github.com/sissbruecker/linkding [2] https://apps.apple.com/nz/app/linkthing/id1666031776
Will probably need to DIY, but if you know any (ideally composable) tools that can cover even a portion of this system, please let me know before I start wiring Buku into Emacs.
1. For Atom/RSS feeds, I use elfeed-org[0]. (including YouTube channels). You can also use elfeed-tube[1] to watch videos on mpv without accessing the website. 2. For spaced-repetition, I also use org-drill[2], this makes it easy to take notes while reading PDFs with pdf-tools[3]. 3. For priority tracking, I use the builtin org-agenda with org-super-agenda[4] to create "views" based on some metadata[5].
This is all done in the same plain text format (org files) and synced in a Git repository.
[0]: https://github.com/remyhonig/elfeed-org
[1]: https://github.com/karthink/elfeed-tube
[2]: https://melpa.org/#/org-drill
[3]: https://github.com/vedang/pdf-tools
[4]: https://github.com/alphapapa/org-super-agenda
[5]: https://gluer.org/blog/2023/simple-org-mode-setup-things-don...
It doesn't collect any information and doesn't require login or registration.
It offers a self-deployed version and provides desktop clients for Windows, Mac, and Linux.
It supports full-text search.
It supports annotating web pages.
There is an open-source and free browser plugin that allows real-time annotation in the browser: https://github.com/hamsterbase/hamsterbase-highlighter.
emacs org-mode + org-cliplink + macro bound to F5. F5 -> insert org-mode style link with URL and Title. Then I can add as much description or tags as needed. If full text is needed, I might pull the site up with eww and copy/extract the text into org entry or add to org-roam for future processing.
New bookmark file per day synced via self hosted Git repo accessed over TailScale.
Manual, but full control and only available to me.
Anybox is a native bookmark manager available in macOS, iOS, and iPadOS.
The data are stored in the Apple’s CloudKit and no personal data is collected.
I don’t have a privacy policy because I haven’t gotten around to creating one yet but yeah, like the other guys say:
- full api access - get all your data out as a zip or json or csv - when you delete something it’s truly deleted, not soft deleted.
When you install it, it works, but upgrading and migrations don't work, so you're stuck with an inaccessible database of bookmarks.
I was very disappointed.
Can't say it's good (a lot of rough edges in the UI) but... it works? And I pushed my FF bookmarks to it without problems. Guess I stick with it for now.
Thanks.
Sadly, pinboard is basically a zombie these days.
Mind you I only signed up a few weeks ago so can't speak to reliability over time.
By "zombie", I mean that it continues to function, but it's a dead product.
Support is often non-existent, and there's been no enhancements for years. If I were looking for the same sort of product today, I'd steer clear of it in favor of something that seemed more like its operator cared.
Send links to yourself.
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
V2 then was even more complicated and undocumented.
Has that improved?
A few reasons why you should check it out: - Saving multiple/all tabs with one click - Ability to open multiple/all tabs with one click to continue previous sessions where you left off - Screenshot preview of pages
Full disclosure: I’m its creator, but I’ll do my best to provide an objective answer.
There are many different solutions to this problem. It really depends on which features are important to you, and which approach makes most sense to your workflow and mindset. Here’s a DB of noteworthy players in this space (in my subjective opinion): https://www.notion.so/natannikolic/21debca9c9e7434c9c61b2996...
Most bookmarks managers, especially (desktop) apps such as Notion, Evernote, Raindrop.io, etc. are limited to saving pages 1-by-1 and usually aren’t able to open multiple links at the same time. They tend to be hybrids between read later and notes apps. They use webpage content images for list-item thumbnails (which are less recognisable than webpage screenshots), and are focused on saving the content (images and full-text transcript), not to effectively work with large quantities of links in the browser where you need them.
On the other hand, most tab managers (and neo-browsers) tend to focus mostly on managing active workspaces and always-on web apps, but neglect the bookmarking and ‘personal knowledge system’ angle of bookmarking. Usually this involves a sort of dock/taskbar inside the browser for switching webapps and a Home Screen dashboard for seeing your tab-groups-like workspaces.
Finally, when it comes to privacy, there is a lot of confusion and misrepresentation. Here’s an example of what I mean: https://better.raindrop.io/feature-requests/p/end-to-end-enc...
Most vendors claim to offer E2E encryption — which just means that the data is sent across using the https protocol and the DB on the server is encrypted. However, this just protects data from unauthorised 3rd party breaches. The vendor itself still has the encryption key to the DB and is able to read users’ application data, i.e. your bookmarks.
If the vendor offers cloud sync, it’s pretty much safe to say your application data is exposed, and subject to how much you trust the vendor not to snoop around or sell it.
Which is why most of us offer local storage only solutions.
However, Tablerone aspires to be the first one to offer a 0-knowledge solution (either through client-side encryption or blockchain storage) whereby we would provide the benefits of cloud storage and multi-device sync, but only the user would have the encryption key to read their data. Thus, not having to trust us, but be able to rely on privacy by design.