Case study
Solaris A/V Engine
Automated quality control for long-form audio-visual content: it watches the video so
humans don't have to — flagging dead air, broken audio, missing segments and
inconsistencies, then exporting a review-ready QC report.
≈650%analysis throughput vs. the previous build
129automated tests (Vitest) green in CI
100%offline-capable after first load (PWA)
The problem
QC of long videos eats hours of skilled work
Reviewing hours-long recordings for technical defects is slow, repetitive and
error-prone. Teams either skip it (and ship embarrassing glitches) or burn analyst
time watching footage at real speed.
The solution — v2 rebuild
What changed from v1 to v2
- ≈6.5× faster analysis. Restructured pipeline with code-splitting and smarter row filtering cut wall-clock time per file dramatically.
- Offline PWA. Installable, works without a connection after first load — field-proof for studio and travel use.
- Exportable QC reports. One click produces a structured report reviewers can archive or share.
- Presets & shortcuts. Repeatable configurations and keyboard shortcuts for power users.
- A/B comparison. Side-by-side runs to validate settings before committing to a full batch.
- Local-first Pro licensing. License validation runs locally — no phone-home, no vendor lock-in, privacy-friendly.
- i18n EN/PT-BR + accessibility pass. Bilingual UI with focus-visible navigation, reduced-motion support and screen-reader semantics.
Deployment options
Cloud today, on-premise tomorrow
Hosted web app
Runs in the browser at solaris-av-engine.vercel.app. Nothing to install; updates are automatic.
On-premise desktop (Tauri)
For studios that need footage to never leave the machine: a native desktop wrapper keeps processing fully local. Available as a custom engagement.
Need someone who ships?
Want Solaris deployed for your team, or something like it built? Talk to me directly.
Start a conversation