I run Claude Code in a side panel most of the day and kept tab-switching to Twitter every time a longer task was running, which meant I'd miss the moment it actually finished and come back ten minutes late. Dropped a Tetris ROM into Standboy and now the Game Boy just sits there doing its thing in the sidebar - I glance at it instead of alt-tabbing away completely, so I actually notice when it goes idle again. Setup was genuinely a folder drop, no config file to fight with.
Tamamon
@mfbz Congrats on the launch. Does SuperContra go with deep work?
@zolani_matebese at the moment Standboy supports only Game Boy games but i was thinking to add support also to other consoles in the future! it's running on RetroArch so techinically it's possible to run a lot of them
The Game Boy as ambient indicator that an agent is running async is a fun design call. Curious about the specific signal: is it binary (something is running, nothing is running), or is the screen actually communicating which agent, which task, how far along?
The visualization choice seems load-bearing to me, because "your agent is doing something" is much less useful in practice than "your agent is on step 3 of the migration and waiting for your token approval." Which side did you land on, and why?
@shanegrant thank you for the interesting questions.
yes actually it attach to your Claude code or Cursor agents hooks in order to determine wether to show or not so it's kind of binary at the moment
i wanted to create something that keeps you in the editor while your agent is running. now it's playing a Game Boy, but it could be something else, the core idea is to keep you in the loop while the agent is doing the work
This is such a clever visual cue for 'deep work.' Does the Standboy character have different animations based on the type of task the agent is currently performing?
@rivra_dev no, but that's a good idea! what were you thinking? like when the agent is using some tools it can show itself working? that's funny
The saves live in a folder you control part is what got me — no cloud dependency
@novamaker01 yes, the idea is that you can export them and store them wherever you want, so you won’t lose your progress
This is a fun way to solve the silent agent problem. Sometimes I am not sure whether my script is stuck or just thinking. Having an interactive agent that shows its state or mood makes it much clearer whether it is actually working or not. Beyond the aesthetic side, does it provide any real debugging feedback if the agent hits a loop, a rate limit, or a code issue that it is unable to solve?
@ritikgupta_01 no, right now it opens a Game Boy where you can play your favorite games while you agent works so that you won't leave the editor. that's a cool idea tho, i was thinking of having some other "mode" instead of just playing
@kr1v definitely, nice idea!
this is the kind of thing that sounds dumb for 5 seconds and then you realize people will genuinely keep it installed for months
also feels like a very “current era” product. humans becoming supervisors for agents means idle time inside editors is suddenly a real UX problem and this weirdly addresses it better than productivity tools do
@abhiram5 exactly! i mean, you can run 10 other parallel agents but to stay focused on what you are doing and avoid too much context switching that's useful