holaOS launches Beta 0.1 with persistent workspaces and sub-agents
Holaboss launched holaOS Beta 0.1 with permanent workspaces, sub-agents, and a dashboard for recurring tasks and review loops. Use it if you want persistent project memory instead of reset-every-run agent chats, though the evidence is mostly launch-thread documentation.

TL;DR
- Holaboss shipped holaOS Beta 0.1 as the first product version of its open-source agent computer, according to _akhaliq's repost of the launch and the linked GitHub repo.
- The core pitch, in hasantoxr's launch thread, is persistent workspaces that keep memory, rules, files, tools, and history instead of resetting context every run.
- Beta 0.1 centers on three product pieces, per hasantoxr's feature list: multi-workspaces, sub-agents, and a dashboard for active runs and review.
- hasantoxr's follow-up frames the payoff as compounding output quality across runs 2, 3, and 10, after the user adds corrections, source rules, tone, and checklists.
You can watch a two-minute walkthrough in hasantoxr's launch post, browse the GitHub repo linked from the thread's try-it-yourself post, and the most concrete use cases in the setup examples are not general chat at all. They are recurring creative and operating loops like a research radar, a content engine, customer feedback triage, and a launch tracker.
Persistent workspaces
The product is aimed at the reset problem: agent sessions that perform once, then forget voice, rules, sources, and checklists on the next run.
The launch thread says each recurring task gets its own permanent workspace with memory, rules, and history. That is a more opinionated framing than the usual chat-with-tools UI, because the unit of work is a long-lived project, not a prompt.
Beta 0.1's three building blocks
Beta 0.1 is described as three parts:
- Multi Workspaces: permanent homes for project context, files, rules, tools, and history.
- Sub Agents: task splitting and parallel execution behind one coordinator.
- Dashboard: a control layer for what is running, what needs review, and what comes next.
That package lands closer to an agent operating surface than a single assistant window.
The workflows it is pitching first
The first suggested uses are recurring workstreams:
- Weekly AI or research radar plus founder brief
- A content engine that keeps one voice
- Customer feedback triage
- Product launch tracking
- Client delivery workspaces
Those examples make the intended audience obvious. holaOS is being positioned around repeatable creative and operating systems, not around one-off prompting.
Run 2 is the product claim
The strongest claim in the thread is not about first-run output. It is that the workspace improves after review, once the user adds instructions like "use these sources only," "keep this tone," and "always include this checklist."
That matters because the beta is selling memory as accumulation. The product claim is that edits become durable operating rules inside the workspace, so later runs inherit them instead of asking for the same setup again.
GitHub beta and walkthrough
The public beta is being distributed through the holaOS GitHub repo, which the thread links directly. The launch materials also include a recorded walkthrough video in hasantoxr's post, which is the clearest evidence that this is already packaged as a product beta rather than just a repo drop.