Jeffrey Li

We just launched holaOS Beta 0.1 — the first product version of what started as our open-source agent computer. I recorded a launch video for the direction we’re building toward: AI teammates that can help with work that unfolds over time, not just one-off sessions. The core problem: most agents are still built for a session. But real work continues next week. Research keeps changing. Content needs the same voice and rules. Customer feedback has follow-ups. Launches have blockers, review points, and the next run. Current agents are impressive in one chat, but they still forget context, lose rules, and make you restart the same work again. A few Beta 0.1 notes around the launch: - Multi Workspaces: Each long-running work-stream has its own context, rules, tools, and history for better organization and focus. - Sub Agents: Handle complex tasks in parallel while the user interacts with a single, centralized coordinator agent for a seamless experience. - Dashboard: Track what’s running, identify what needs review, and see the next steps—fully customizable to fit your workflow. The Open Agent Computer is still the foundation. But the user-facing unit is becoming clearer: not a disposable session, not a blank agent builder, but a living workspace for work that unfolds over time. Still early. Try it with one recurring work-stream you keep restarting every week.

@rion978 Sounds like a good business to me

@JeliPenguin Welp, guess I’m selling a course now lol.

We just launched holaOS Beta 0.1 — the first product version of what started as our open-source agent computer. I recorded a launch video for the direction we’re building toward: AI teammates that can help with work that unfolds over time, not just one-off sessions. The core problem: most agents are still built for a session. But real work continues next week. Research keeps changing. Content needs the same voice and rules. Customer feedback has follow-ups. Launches have blockers, review points, and the next run. Current agents are impressive in one chat, but they still forget context, lose rules, and make you restart the same work again. A few Beta 0.1 notes around the launch: - Multi Workspaces: Each long-running work-stream has its own context, rules, tools, and history for better organization and focus. - Sub Agents: Handle complex tasks in parallel while the user interacts with a single, centralized coordinator agent for a seamless experience. - Dashboard: Track what’s running, identify what needs review, and see the next steps—fully customizable to fit your workflow. The Open Agent Computer is still the foundation. But the user-facing unit is becoming clearer: not a disposable session, not a blank agent builder, but a living workspace for work that unfolds over time. Still early. Try it with one recurring work-stream you keep restarting every week.

@AriyanA32556 Of course! Fully open sourced, Fully customizeable!

@JeliPenguin Can it connect to our own model gateway?

@nolancarterai Yes of course! Reach out to us if youre interested!

@JeliPenguin Will your team handle implementation and onboarding?

@setu_ai_expert Of course! And it's already patched! Checkout out latest release

@JeliPenguin Used this for a week and my CPU hit 114%. Any optimization coming? 😭

@pumpkherm We feel you — humans should enjoy their tools, not endure them.

@JeliPenguin I’m so tired of OpenClaw / Hermes-style CLI UX. Finally, a super-agent with an actually nice interface.

@SuccessSparksX Marketplace is already a work in progress! creators deserve a stage!

@JeliPenguin Is there gonna be a Workspace marketplace? If I build one, can I sell somewhere?

@ethan_walker_ai We just might be… bonus points if you can explain AI to your grandma.

@JeliPenguin Y’all hiring sales?

@CodeByAurelia Exactly. Survival > theatrics. We want agents that stay alive with your project.

@JeliPenguin A lot of browser agents are still demoing button clicks. This is already talking about how a project actually survives.

@altwba The messy part is real; our mission is to make that invisible to users.

@JeliPenguin The concept is right. The hard part is making sure it doesn't get messier the longer you use it.

@Nijol71 Humans + agents = superpowers. The office version of the Avengers.

@JeliPenguin Future bosses won’t win by hiring more people. They’ll win by knowing how to use this.

@Jubayer7792821 Writing tools are cute, but we’re trying to run marathons, not sprints.

@JeliPenguin This is already playing in a different league from AI writing tools.

@SphereSuc32514 Each strategy gets its own space to grow. Chaos-free investing, hopefully

@JeliPenguin One Workspace per strategy? Yeah this is kinda perfect for investing.

@ExpertKelvin_AI If GUI moments are this rare, we’ll take it and run.

@JeliPenguin This feels like the Mac GUI moment for agent software.

@4111J_ We’re trying to make AI feel less like a tool and more like a teammate.

@JeliPenguin Coze feels more like orchestration. This feels more like a collaboration environment.

@SuccessBluePrtX That’s the dream: built to remember, organize, and keep moving, not just respond.

@JeliPenguin A Workspace is basically an AI-native app.

@atulkumarzz Glad it resonates! One long-running work-stream at a time, we hope to make it the new normal.

@JeliPenguin This is incredibly well-articulated and genuinely exciting. The concept of an "Open Agent Computer" evolving into a living workspace for unfolding work is exactly what the industry needs.

@Mohibur_Tech_Ai Workspace is the polite term… our agents would just call it home.

@JeliPenguin Calling it a workspace might actually be underselling it.

@Adilson_Ai Roadmap exists! Still figuring out if it should come with a cape or just really good UX.

@JeliPenguin Is there a roadmap for the admin console?

@LunarResearcher And you, my friend, are part of it

@JeliPenguin holaos is our future

@hey_dinislam Yep! One chat is fun; multi-workstream continuity is next-level.

@JeliPenguin ChatGPT is good at answering. holaOS looks like it can actually keep working.