@droid Please pay attention to this issue; this bug has existed for a long time, and we hope it can be fixed as soon as possible.
https://t.co/zmmKumGmiC
I found the winner the level of polish makes me happy.
Support Droid, Hermes, Codex, Pi, and much much more. Clean design, uses terminals as it should.
https://t.co/dxugGzqN2Q
Our official skills repo is open source: https://t.co/X3yOgU0tRd
Equip your agents with curated skills for iOS and Android development, Office file editing, and visual effects with GLSL shaders.
There are more open source projects coming!
My Conductor workflow:
- new workspace (⌘N) for every bug/feature
- spin up the dev server to test (add a run script so ⌘R runs server for you)
- if all looks good, create PR (⌘⇧P)
- kick off a review (⌘⇧R), look at diffs + leave comments
- merge, then archive (⌘⇧A)
- start again!
Anthropic shipped generative UI for Claude. I reverse-engineered how it works and rebuilt it for PI.
Extracted the full design system from a conversation export. Live streaming HTML into native macOS windows via morphdom DOM diffing.
Article: https://t.co/BcYo94YqK3
Repo: https://t.co/EfMDX58NWc
Built on @badlogicgames's pi and @DanielGri's Glimpse.
https://t.co/4Fo3NK200e is insane.
Note: I’m not affiliated with Pi, or the subagents extension mentioned below (yet) 😄, just genuinely amazed by this tech.
I won’t hit you with the programmers out of jobs meeeeeeh
But I will tell you that it’s super fast, ships its own docs so your AI can build plugins for you, and some of the community plugins are insanely useful.
Example: https://t.co/t32TYrg8qN by @nicopreme
Run subagents in parallel, create chains where they run one-by-one while sharing info with the next one, or mix both.
Had my ai come up with some slash commands like /spike that takes my prompt and then thoroughly researches the codebase & internet docs to come up with a grounded plan for whatever I want to build.
Also trying a /deliver command that does research, builds plan, implements plan, writes tests, and then reviews its own code.
It created an extension for me that imports the subagent extension and calls its functions.
It’s insane how hackable this thing is.
I’ve tried Claude code, codex CLI, factory, opencode
And I always come back with
Man I wish it had feature X from the other ones.
Or man, I wish this worked a bit differently than it does today.
Heck, I built a swarming skill for codex that uses the MCP for sub agents but ran into issues because it could only call MCP tools sync so no parallel agents without building a planner.
With Pi, I just tell gpt5.4 what I want and it goes and builds or updates extensions for me.
I also love their ethos:
> There are many coding agents, but this one is mine.
Makes me think of brands that apply the “you” branding where they let you customize their services to your liking, to make it “you”.
In a world that is increasingly run, and limited by robots, I like it when tools focus on the human aspect. Powered by robots to unlock new possibilities, but uniquely tailored to 𝘆𝗼𝘂, the human.