Discussion

Search code, repositories, users, issues, pull requests...

billconan: does it use the claude code api or the claude code cli? You know, the claude code api is more expensive.I also hope it can have a webapp version, rather than electron. because most of our work are on a remote server.

ZeidJ: It actually uses the Claude code SDK so it plugs into whatever you already have.It can use API/CLI or even if you have a private hosted instance.We're actually working on a remote web app version but its a little trickier to wire up.These are great questions - thank you!

linsys: If it uses the SDK then it's token burn? Or can it "legally" use your Claude.ai MAX account, your subscription account?

ZeidJ: Great question - It can "legally" use your subscription account! We wanted to make sure it was more accessible.

james-clef: Maybe a bit of an odd one, but can you decorate the office? I'm wondering like have you abstracted the decor elements into something that is straightforward to extend? How easily could I give my office a new espresso machine or something?

Ryand1234: this seems interesting. Will give it a spin this weekend.

ZeidJ: Thank you - let us know if you have any feedback. Appreciate you trying it out.

lanxevo3: The parallel group + sequential handoff pattern is how we run our agent fleet too. One thing we found that helps with the "when to bother a human" problem: let the model output a confidence score with each decision. Below a threshold, it asks. Above, it proceeds. Over time you tune the threshold based on how often the model was right vs. wrong on auto-decisions. Works better than a fixed deny/allow list for the long tail of ambiguous cases. The pixel office visualization is a great idea — makes the orchestration state legible to a human overseer in a way terminal logs never do.