Is this project still active?

[dfehring-thezenith]

Love the project, just wondering if it's abandoned and I should make a fork, or if it's just that the main dev is busy/on vacation/etc. Haven't seen any activity in a while

[hmans]

Hi! Yes, it is, and I'm still here, just super super busy with another Project at the moment.

[hmans]

PS. feel free to make a fork, OSS is beautiful!

[hmans]

https://www.hmans.dev/blog/beans-update

[YAMLcase]

Of all the llm task managers I've tried, I think this is the best! I hope it sticks around

[hmans]

I'm so happy to hear you folks dig it! I'm trying to make up my mind where to go with this next. My focus at the moment is almost exclusively on my big project Chatto https://www.hmans.dev/blog/chatto-is-open-source and will need to stay there for a little while longer.

Personally, I've never used the Beans TUI much myself, so I'm facing a bit of a paradoxon: the actual task tracking functionality that Beans provides can probably, at this point, quite easily be replaced with a good agent skill. I understand that people also love the TUI, but it's the part of Beans that I care about the least.

So... I don't really have any news at this point. But it's likely that when I find time to invest into Beans, there's a chance I'll supersede it with a separate tool to replace it.

If anyone wants to take over maintenance of the project, hit me up and we'll talk!

[YAMLcase]

just scrap the TUI! beans is awesome because if it's simplicity and there's no other game in town like it. with something like beads others are writing the UI not because the author asked, but because people wanted it.

[Antriel]

I use the TUI because I don't want to waste tokens asking the agent to tell me what we were doing. I even constantly type in the IDs manually. But the TUI sucks, I would rather like the kanban UI idea pushed further forward.

I also like the idea of turning it more into a skill, and more focused on direct markdown file editing. That's what the models are trained for, there's not much point in replacing that with API. But the search/discovery and general management, that could be useful to have some API over it. It's hard to say what kind of API would work best for all this. Agent replacing whole lines just to mark a checkbox, often having shell issues eating backticks and such; all that could have simpler interface for the agent maybe. Not sure.

Another thing I would like is to not force the agent to use beans all the times, I find myself often adding "no need to create a bean for this task" in my prompts.

But mainly the web UI, I want to see my beans in better view, see dates, easily edit, group, tag, color...