PictaBase — A pixel-blind, relational visual database for teams

I built PictaBase to solve the "dumb folder" problem in high-stakes visual workflows. In film production, finding a specific continuity shot across thousands of files is an operational bottleneck. This platform treats photos as relational data, not just files.

The Technical Footprint (May 2026)

This is not a thin wrapper. It is a professional-grade application built with a strict quality gate:

• Backend: 287 PHP 8.4+ files (38,500 lines) enforced at PHPStan Level 8.

• Frontend: 131 TS/TSX files (26,967 lines) using React 19 and Vite 7.

• Database: 17 custom tables managed via dbDelta() with synchronous schema-version locking.

• Coverage: 488 PHPUnit tests and 1,741 assertions pinning the business logic.

Why it’s different for Makers

• Pixel-Blind Architecture: Your image bytes never transit the PictaBase server. We use S3 Presigned POST for direct browser-to-bucket uploads.

• Sidecar Metadata: Tags, notes, and AI labels are written to .meta.json files alongside the originals in your bucket. If the database disappears, your data survives in standard JSON.

• Fail-Closed CDN: Egress is managed via CloudFront with project-scoped signed cookies. If CloudFront signing is unavailable, the system fails closed (503) rather than silently falling back to expensive S3 fetches.

Test Drive the Free Tier

We are offering a fully functional Free Tier (250 MB) with zero feature gating. You can stress-test the S3 pipeline, the faceted search, and the AI suggestions immediately without a credit card.

I am a solo developer. If you find a way to bypass the rate-limiter or break the 4-code hard cap in LtdCheckoutIntentController.php, please report it through the support portal.
@PictaBase