This is very cool! Right now I have all sounds turned off but that makes me keep an eye on windows, which I really don't enjoy. Love this idea of sounds for states and identities. Makes me want to have it say "Hey, can I get your eyes on this?" instead of a ping.
Thanks. I was window watching, or window forgetting, too much. It led me down this road. I used to call my Apple watch my “shock collar” because I needed something to startle me into noticing I needed to do something. I’ve found I can still tune out the GitHub voice notification if I’m in the zone, but overall I’ve found it super helpful.
Using the GITHUB_ACTIONS variable to give the runner its own voice is a clever fix. "Lovelace is done" tells you who finished without looking at a screen, which matters a lot with four projects open. I also liked the warning that Codex allows only one notify line, since that one would break another feature quietly. Is Sully getting a spoken name next?
Sully sounds like the tricky one, since Lovable runs in a browser tab instead of VS Code. Watching for its commits on the GitHub issue might be a way in, since you already point Sully at an issue. Is Wiggins another agent in the rotation too?
Thanks for the question. I currently use Sully via Slack with all the others. Yes Wiggins is another. I am curious if I can leverage that pathway. Direct Sully to read the issue and act. That's all the direction I give now since I have all the work done in Lovable via issues. Lovable also has an app and I was considering if I could leverage playwright. So many fun things to try. Prioritization is my problem. I am also excited to try this with a local model or two. Maybe save some money. We will see what I come up with.
Directing Sully to read the issue and act is a neat setup, since the issue becomes the whole brief. Playwright on the Lovable app sounds like a fun one to try. With this many options, I get why prioritizing is the hard part. Which one are you trying first, the local model or Playwright?
This is very cool! Right now I have all sounds turned off but that makes me keep an eye on windows, which I really don't enjoy. Love this idea of sounds for states and identities. Makes me want to have it say "Hey, can I get your eyes on this?" instead of a ping.
Thanks. I was window watching, or window forgetting, too much. It led me down this road. I used to call my Apple watch my “shock collar” because I needed something to startle me into noticing I needed to do something. I’ve found I can still tune out the GitHub voice notification if I’m in the zone, but overall I’ve found it super helpful.
Ha! 😂 . You keep posting things I need to setup and I still haven’t had time!!
I relate to this problem! My list keeps growing as well. I need more hours in my day.
Or an agent to manage it? 😂
Bookmarked :)
Using the GITHUB_ACTIONS variable to give the runner its own voice is a clever fix. "Lovelace is done" tells you who finished without looking at a screen, which matters a lot with four projects open. I also liked the warning that Codex allows only one notify line, since that one would break another feature quietly. Is Sully getting a spoken name next?
Thank you. Determining if I can do the same with Sully and Wiggins is on my todo list.
Sully sounds like the tricky one, since Lovable runs in a browser tab instead of VS Code. Watching for its commits on the GitHub issue might be a way in, since you already point Sully at an issue. Is Wiggins another agent in the rotation too?
Thanks for the question. I currently use Sully via Slack with all the others. Yes Wiggins is another. I am curious if I can leverage that pathway. Direct Sully to read the issue and act. That's all the direction I give now since I have all the work done in Lovable via issues. Lovable also has an app and I was considering if I could leverage playwright. So many fun things to try. Prioritization is my problem. I am also excited to try this with a local model or two. Maybe save some money. We will see what I come up with.
Directing Sully to read the issue and act is a neat setup, since the issue becomes the whole brief. Playwright on the Lovable app sounds like a fun one to try. With this many options, I get why prioritizing is the hard part. Which one are you trying first, the local model or Playwright?