May 20, 2026

KaKeKiKoKu Studio for Linux is here

The promised Linux build is now available as an Early Access Snap Store Candidate.

Update (July 2026): KaKeKiKoKu Studio for Linux has since moved to the Snap Store's stable channel — install with sudo snap install wuniq. It is still Beta; only the delivery channel changed. The Early Access note below is kept as a record.

In the KaKeKiKoKu Studio 1.2 post, I said the next step was Linux. In the KaKeKiKoKu Studio 1.3 post, I said the Linux build was finished and waiting for its public door.

That door is open now.

KaKeKiKoKu Studio for Linux is available as an Early Access Snap Store Candidate. Not stable yet, deliberately. Studio itself is still in Early Access, and this first Linux publication should keep that shape: public enough to use, installable enough to test properly, cautious enough not to pretend the journey is finished.

sudo snap install wuniq --candidate

The Snap includes the Knowledge Engine, the web UI, and the MCP entry point used by AI coding clients:

/snap/bin/wuniq.wuniq-mcp

What has been validated

This is not a theoretical port. Studio has been tested on Linux with the three current AI paths I care about most: Claude Code, Codex CLI, and Antigravity. That covers Anthropic, OpenAI, and Google workflows without turning Studio into a tool owned by any one of them.

Claude Code remains the only client with Studio channels today, so it can run in Automatic mode. Codex CLI and Antigravity work through Precise mode: Studio provides context through MCP tools, and UI actions that cannot be pushed directly to the AI are prepared for the clipboard. That distinction matters, but it is not a hierarchy. Studio is designed to be useful in both flows.

One Linux note

If your projects live under your home folder, the Snap should work normally. If your projects live in /mnt, /media, a USB drive, or a virtual-machine shared folder such as /mnt/hgfs, Snap's sandbox needs one extra permission:

sudo snap connect wuniq:removable-media

This is not a Studio bug. It is normal Snap confinement. After granting the permission, reloading the Studio web UI is enough.

Shared folders and network mounts can also miss file-change notifications on Linux. Studio still opens the project, but automatic change detection may not fire reliably there. In that case, run sync from your AI client or click Analyze in Studio.

Why Candidate?

Because this is the first public Linux build. Candidate is the honest channel: it says the work is real, the package is public, and users can try it, while still leaving room for the first wave of Linux-specific reports.

When the Linux build has lived in public for a while, I can promote it to stable. Until then, this is the right shape for KaKeKiKoKu Studio: available, useful, and still careful.

Studio started on Windows because that is where I live day to day. But the engine was never meant to be trapped there. It is a local Knowledge Engine for AI work, and Linux is a natural home for that kind of tool.

So this is a small post, but a meaningful one. Studio now has a Linux path.