media-upload

name	media-upload
description	Upload scored media to Vercel Blob or Supabase Storage by quality tier
trigger	pipeline
depends_on	["taste-score"]
sandbox	false
inputs	[{"name":"hero_threshold","type":"int","description":"Minimum TASTE score for Vercel Blob upload","source":"config.yaml#arcanea_claw.upload.hero_threshold"},{"name":"gallery_threshold","type":"int","description":"Minimum TASTE score for Supabase Storage upload","source":"config.yaml#arcanea_claw.upload.gallery_threshold"}]
outputs	[{"name":"uploaded_count","type":"int","description":"Number of assets uploaded this cycle"}]
dependencies	{"python":["supabase"],"node":["@vercel/blob"]}
tables	["asset_metadata (UPDATE storage_url, thumbnail_url, storage_tier)"]

media-upload

Routes scored assets to the appropriate storage backend:

• hero (80+) — Vercel Blob via Node subprocess
• gallery (60-79) — Supabase Storage bucket arcanea-gallery
• thumbnail (40-59) — Supabase Storage bucket thumbnails

Assets below 40 (reject tier) are not uploaded.

Pipeline Position

Step 6 of 8 — runs after taste-score, feeds social-prep.