what I'm building now: August 2026
A working note on the AkurAI services, BifrOSt, and rebuilding this site while the portfolio grows.
I have 28 projects marked in progress in Multica. That number is less useful than it sounds: it tells me the portfolio is broad, not what I am doing today. The work with recent build evidence is narrower—AkurAI Platform and Router, BifrOSt, and this site.
AkurAI: integration work, not another app list
AkurAI Platform is still the integration point for the self-hosted services. The project cards already list Mail, identity, VPN, Drive, Notes, CRM, Router, Monitor and DNS, so I will not repeat that catalogue here.
The useful status is in the build history. Four Platform revisions built successfully on 8 August. A newer revision is currently waiting in the pipeline. Router also had a successful build on 8 August. That is where this work sits: individual services exist, but the current job is keeping their shared gateway and delivery path coherent.
BifrOSt: distribution work is moving separately
BifrOSt is not a side feature of AkurAI. It is its own Arch-based workstation and currently has separate repositories for the distribution, applications and website. The distribution's latest recorded run succeeded on 8 August, while the Apps and Web projects remain tracked independently.
That split costs more coordination, but it keeps OS release work, application experiments and public documentation from pretending to share one lifecycle. For now I prefer that honest separation over one repository with a misleading single status.
This site: replacing placeholders with evidence
The most recent shipped work is this site itself. On 9 August I changed it from a thin demo into an editorial portfolio, added BifrOSt to the public project list, marked the placeholder posts as drafts, and wired deployment to follow a green build. The latest site run succeeded.
This post is part of the same cleanup. I would rather publish a short note tied to current project state and build runs than keep generic copy about what I might write later.
The tradeoff
The blocker is coordination, not a missing idea. Multica currently tracks 96 issues across the 28 in-progress projects, with 85 already marked done. A green run proves one revision passed its pipeline; it does not prove the portfolio is finished or that every service should move at once. The newest Platform run is a useful example: it is waiting, while the previous revision succeeded.
So I am keeping the next slice narrow: finish integration and delivery work where runs show active movement, keep BifrOSt release work separate, and make this site report what actually shipped. The project count can stay large. The work in flight should not.