When I found cool sites, or did a summarize, i feel the urge to hit "fav", "like" or "bookmark". Maybe it´s nice to have rudimentary bookmark feature in kagi?
I know this is already marked as Planned, but I wanted to jump in and +1 this suggestion... I use Raindrop.io (though have searched far and wide for other Bookmarking service options). I would love a Kagi Bookmarking service. Let me bookmark sites that are returned in searches and additional sites via a Kagi browser extension.
And then give me a Kagi Lens to search Bookmarked sites only. This would be a hugely useful service for me to collect bookmarks for research, documentation, product manuals, etc.. and then be able to use that as an easily referenced knowledge repository.
While this does have potential privacy considerations. Either make it clearly opt-in where a user accepts awareness that this is storing user data with a reminder that Kagi will not share or sell any user data OR end to end encrypt the bookmark data using the user's Kagi password to decrypt in-browser (think Proton Mail). A Lens should still be achievable by having the browser pass the full list of decrypted bookmarks to the search.
invisible Fun fact, we have a bookmarking service at https://tinygem.org but it is not integrated into kagi. But we hear you, we'd love to build a minimum version of result bookmarks.
Do you have any ideas in terms of UI?
How do you bookmark a site? (how many clicks to do this?)
How do you save it to a folder or tag it while doing it?
How does bookmark management look like?
How other feature look like?
Ideas, mockups, screenshots would help us understand the goals important to you and what a minimum version should have. cc @polarity
For me, it'd be nice if Kagi could be used as a mix of a private bookmarking service (like Pinboard or Raindrop) and a private search engine (like YaCy).
I save every interesting article, service, etc that might come in handy in the future. So now I have 10k+ bookmarks in Raindrop. But in 80% of cases when I need to find something, I get dozens of results but not the one I need.
The regular bookmarking services are more suited for organized links, where you tag them and add some description. But if you just want to save links with one click (e.g., via browser extension) then it'd be tricky to find them later if you have a lot of links.
Search engines like YaCy do a bit better job with the search, but adding new links requires a lot of clicks.
So if Kagi can combine the best of both worlds (ease of adding new links + awesome search engine), it'd be nice.
I'm seeing the conflict of goals concerning privacy... nonetheless it would be great being able to narrow the search to sites I bookmarked already. As of now I can only search my bookmarks by the site's title or address.
Well, it's just the ... bookmarks, i.e. the sites I added to the bookmark list in my browser.
Suggestion: Add the ability for users of Pinboard to connect Kagi to their Pinboard account via API, and pull results from the the user's saved Pinboard items into their Kagi search results.
There is probably a large overlap between Pinboard and Kagi users, and Pinboard users save items in Pinboard for later reading or to create a knowledge repository. However, having to search in both Pinboard and Kagi is cumbersome, and most people will default to searching Kagi — while neglecting to search the data they're already curated. Pinboard offers full-text indexing and archival of saved items, which means it would be cost-free for Kagi to query and get all results containing the searched keywords. Pinboard would handle getting the relevant saved items; Kagi would only need to display them.
If a user could search through Kagi and have their saved Pinboard items pulled in (ideally, clientside with no communications to Pinboard happening on Kagi's servers!), that would make Kagi a central place for the most relevant search results to that user.
In terms of the UX, I would expect there to be a section for "Pinboard Integration" in Settings > Personalized Results. The user would be able to enter their Pinboard API key (and, optionally, select if they want to only search public or private saved Pinboard items. [Public saved Pinboard items are listable via RSS feed]).
Then, when the user makes any search on Kagi, a clientside request would be sent to Pinboard's API to find any results containing the query (Pinboard would handle the indexing and searching work of finding the query in a saved item's title, description, or, if archived, full-text). Pinboard would then return the relevant results for display at the top of Kagi search results. The Pinboard results could be lazy-loaded in. They would be displayed at the top, as they would be the most relevant to the user due to having been manually curated and saved by them.
I would be willing to fund a bounty on this.
Relevant Links:
An alternative direction to consider: Allow archiving/saving items directly in Kagi itself. For example, if I read an awesome blog post, I can "save" it to Kagi, have Kagi full-text index it (and maybe submit it to Internet Archive), and then prioritize it in my search results.
Basically, build Pinboard-like functionality directly into Kagi to make it so Kagi can become a person's knowledge repository.
I'd say that we can understand if there's a user need about this integration type.
For example, I was a former Pinboard user, but due to some medium/long term consideration (and archiving capabilities), now I host my 20k+ bookmarks on https://raindrop.io
Raindrop.io has also a stable API (https://developer.raindrop.io/) , a token-driven integration, etc
So I think that the main question boils down to this: is valuable for Kagi to integrate with bookmarks manager?
Does it add lots of complexity, achieving a small benefit...or not?
How can we formalize Kagi ranking of bookmark result? Will they be automatically promoted? How much?
I think that all users would recognize visually a search result pulled from the bookmark manager, but how we'll do it?
It would be better to start having only a lens for that?
These are only few questions, I'm trying to share here that in my opinion this is quite a complex proposal; obviously I'd be very happy to see raindrop.io integration, but we have to share a mental model about that.
seventhwave I might switch over to Raindrop, especially with the recent issues Pinboard has been having. Thanks for the link!
kodex (then I'll stop talking about Raindrop in a Kagi forum) very happy user here; has great responsiveness (has closed 2 bugs I've opened), more feature, more effectiveness. and archiving here works well (not so in Pinboard). the fee is, as for Kagi, well deserved and essential to product sustainability.
This is in principle a great idea, the problem arises from the fact that the adoption of these tools is very fragmented. Pinboard/raindrop/pocket are just some of many tools used for this and integrating them means spending a lot of time on each and then still not having all of them covered.
Instead we should probably build bookmarking tool inside Kagi itself and allow people to import a bookmark file to it as well as add links through the interface.
What do you think?
Vlad Thinking about Kagi-the-product, I think this could be a good area to explore.
However, I'm a bookmark power user: I need archiving, mobile, tagging, searching, etc.
The real divider for me is archiving: if you try to go into the Pocket territory (being a tool with tagged bookmarks), I think would be feasible for Kagi, and maybe we can stay with the current monthly fee.
Instead, if you're thinking to go deep and offer features like Raindrop (archiving, full text search, a mobile app, highlighting, sharing), you've to calculate the right fee (I think that the annual Raindrop fee is not only fair but also quite optimized being specialized, IMHO)
In my opinion try this would be a winning move for us all: this could tightly integrate result search and bookmarks, and who don't want that? However, for power user like me, I'd need a strong offer on Kagi side to migrate from Raindrop.
