I’ve been working on something super cool.
Since building @AgentKeysApp and @PredictwithOshi, my biggest issue has been how dispersed my AI workflows are.
- Codex
- Claude Code
- 9 @OpenClaw Instances with 6-12 agents per instance
- My Hermes Agent
- 20+ API keys floating around
Most of my conversations are in different chats across different applications, and some of them are even stuck to desktop conversations.
And in order for them to be in a workflow, I still need to be the middleman, copying and pasting between apps and LLMs.
So I’ve built @aOSWorkspace.
aOS is designed to replace scattered AI tabs with one surface where chats, Codex, Claude Code, voice, files, approvals, and workflows all move through shared workspace state.
All via a mobile app, web app, and desktop app.
Connect your API keys, agents, or even a Codex/Claude Code chats into 1 Operation system.
Organize them into specific workflows, allow allowing multiple LLMs to work on one project at the same time, and much more.
aOS is full, one stop operating system for all your AI needs.
Then, connect to the aOS Agent App Store: download, operate, or even create applications for AI agents to run autonomously through the aOS operating system.
Localization is becoming more and more important for AI agents, and this is something I originally built solely for personal use.
But I’m considering to make it public as it integrates perfectly with https://t.co/guDiFl19i9.
The future is agentic.
And humans need an operating system for their agents.
Public GitHub will be here when live: https://t.co/am4hyCQgEW