[GH-ISSUE #384] [feature request] trigger webhooks #252

Closed
opened 2026-03-02 11:48:03 +03:00 by kerem · 7 comments
Owner

Originally created by @4383 on GitHub (Sep 10, 2024).
Original GitHub issue: https://github.com/karakeep-app/karakeep/issues/384

Hey there,

I'm looking for a bookmark solution that allow me to trigger webhooks to a specified url once I bookmark a thing (a note, an article, a blog post et....).

AFAIK this is not feasible with hoarder.
Do you think it would be possible to implement this kind of feature?

Thanks for all the work you have already done.

Originally created by @4383 on GitHub (Sep 10, 2024). Original GitHub issue: https://github.com/karakeep-app/karakeep/issues/384 Hey there, I'm looking for a bookmark solution that allow me to trigger webhooks to a specified url once I bookmark a thing (a note, an article, a blog post et....). AFAIK this is not feasible with hoarder. Do you think it would be possible to implement this kind of feature? Thanks for all the work you have already done.
kerem 2026-03-02 11:48:03 +03:00
Author
Owner

@MohamedBassem commented on GitHub (Sep 15, 2024):

It's indeed currently not possible. Do you mind expanding a bit on how you'd use this?

<!-- gh-comment-id:2351236341 --> @MohamedBassem commented on GitHub (Sep 15, 2024): It's indeed currently not possible. Do you mind expanding a bit on how you'd use this?
Author
Owner

@4383 commented on GitHub (Sep 17, 2024):

Sure.

Here is my personal use case:

I daily read books, during my reading I take notes and store links into hoarder.
I store my personal notes by using the obsidian markdown editor to create my own zettlekasten (an obsidian vault).

But today to automatize my note taking process, I created my own system:

  • I launched a local instance of the llama 3.1 LLM.
  • I connected my llama to a telegram bot, and, when during my reading, I found a useful concept, I send the concept name as a message to my telegram bot, and then, this bot trigger a llama prompt
  • once the LLM prompt is done, the bot create a new note in my obisidian vault. A note related to the concept I passed as message and built with the output of llama.

My knowledge base is growing automatically in parallel of my reading and I have no friction due to note taking during learning.

If I'm able to simply store in hoarder the link of web page explaining the concept, by example a wikipedia page, and then trigger web hook to my local LLM endpoint, then I'd be able to remove lot of mechanisms I used to create this workflow.

The ideal workflow would be:

  1. I create a bookmark;
  2. the LLM prompt is triggered by the hoarder webhook on bookmarks additions (by passing the title of bookmark);
  3. my note is automatically created in my vault on the result of the LLM prompt request.

My LLM endpoint would be exposed by a script to manage the IO operations (the local file creation).

Let me know if you need more details.

<!-- gh-comment-id:2355078807 --> @4383 commented on GitHub (Sep 17, 2024): Sure. Here is my personal use case: I daily read books, during my reading I take notes and store links into hoarder. I store my personal notes by using the [obsidian markdown editor](https://obsidian.md/) to create my own [zettlekasten](https://en.wikipedia.org/wiki/Zettelkasten) (an obsidian vault). But today to automatize my note taking process, I created my own system: - I launched a local instance of the llama 3.1 LLM. - I connected my llama to a telegram bot, and, when during my reading, I found a useful concept, I send the concept name as a message to my telegram bot, and then, this bot trigger a llama prompt - once the LLM prompt is done, the bot create a new note in my obisidian vault. A note related to the concept I passed as message and built with the output of llama. My knowledge base is growing automatically in parallel of my reading and I have no friction due to note taking during learning. If I'm able to simply store in hoarder the link of web page explaining the concept, by example a wikipedia page, and then trigger web hook to my local LLM endpoint, then I'd be able to remove lot of mechanisms I used to create this workflow. The ideal workflow would be: 1. I create a bookmark; 2. the LLM prompt is triggered by the hoarder webhook on bookmarks additions (by passing the title of bookmark); 3. my note is automatically created in my vault on the result of the LLM prompt request. My LLM endpoint would be exposed by a script to manage the IO operations (the local file creation). Let me know if you need more details.
Author
Owner

@ebenoist commented on GitHub (Nov 24, 2024):

I would love to see this or just an RSS feed of my hoarded items. I had been building a similar application to hoarder before I discovered it and quickly abandoned my plans of building my own (thank you). Here's the code I used to set up a simple RSS feed based on the items I've saved.

If you're interested I could probably port this functionality to hoarder itself.

https://github.com/ebenoist/enlace

and here's an example of a rendered feed: https://enlace.space/~erik/rss.xml

The nice thing about RSS is it provides a simple machine readable interface for collections like links, which hoarder does a great job in providing in app, but doesn't share that common interface.

<!-- gh-comment-id:2496228628 --> @ebenoist commented on GitHub (Nov 24, 2024): I would love to see this or just an RSS feed of my hoarded items. I had been building a similar application to hoarder before I discovered it and quickly abandoned my plans of building my own (thank you). Here's the code I used to set up a simple RSS feed based on the items I've saved. If you're interested I could probably port this functionality to hoarder itself. https://github.com/ebenoist/enlace and here's an example of a rendered feed: https://enlace.space/~erik/rss.xml The nice thing about RSS is it provides a simple machine readable interface for collections like links, which hoarder does a great job in providing in app, but doesn't share that common interface.
Author
Owner

@MohamedBassem commented on GitHub (Nov 24, 2024):

@ebenoist Producing an RSS feed is tracked in #333. To be honest, it's currently on not my high priority list, but I'd be open to accepting a contribution if you're up for it :) You can join us on the #development channel in discord to discuss the specifics if you're interested.

<!-- gh-comment-id:2496312839 --> @MohamedBassem commented on GitHub (Nov 24, 2024): @ebenoist Producing an RSS feed is tracked in #333. To be honest, it's currently on not my high priority list, but I'd be open to accepting a contribution if you're up for it :) You can join us on the #development channel in discord to discuss the specifics if you're interested.
Author
Owner

@ebenoist commented on GitHub (Nov 24, 2024):

@MohamedBassem I'm down to help out. I don't think it'd be difficult. I have the basics, I just need to rewrite it in typescript. I'll jump in the channel if I have any questions!

<!-- gh-comment-id:2496319738 --> @ebenoist commented on GitHub (Nov 24, 2024): @MohamedBassem I'm down to help out. I don't think it'd be difficult. I have the basics, I just need to rewrite it in typescript. I'll jump in the channel if I have any questions!
Author
Owner

@MohamedBassem commented on GitHub (Jan 19, 2025):

This got implemented in #852 and will be available in the next release.

<!-- gh-comment-id:2600985280 --> @MohamedBassem commented on GitHub (Jan 19, 2025): This got implemented in #852 and will be available in the next release.
Author
Owner

@4383 commented on GitHub (Jan 19, 2025):

Thanks! Much appreciated

<!-- gh-comment-id:2601001944 --> @4383 commented on GitHub (Jan 19, 2025): Thanks! Much appreciated
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
starred/karakeep#252
No description provided.