fluxtube.forklabs.cc
FluxTube
A serverless sync job that bridges Miniflux with YouTube playlists. Reads unread YouTube entries from your configured Miniflux categories, adds them to the playlists you map, and marks them read once you've watched and removed them on YouTube.
What it does
You already use Miniflux for RSS. YouTube channels publish RSS. You watch on YouTube — for the offline downloads, the cross-device progress sync, the TV app. The missing piece is getting unread items into a YouTube playlist without copying URLs by hand. That's FluxTube.
Categories map to playlists. The Worker runs on a cron tick (every 30 min by default), adds anything new, and watches the playlist for what you've removed. Removed = watched = marked read in Miniflux. The loop closes.
Stack
- Cloudflare Workers · cron-triggered TypeScript
- Cloudflare D1 · per-entry tracking state
- YouTube Data API v3 · playlist reads + inserts
- Miniflux REST API · category reads + bulk mark-read
- Healthchecks.io · dead-man's-switch + typed failure routing
- Grafana Cloud · structured logs (Loki) + metrics (Mimir / OTLP)
- Terraform · everything; release-please owns versioning
Run your own
FluxTube is single-tenant by design — no shared infrastructure, no accounts, no signups. The whole sync is your Cloudflare Worker running against your YouTube account, your Miniflux, your Healthchecks. The repo's TODO.md walks the bootstrap top-to-bottom; about an evening end-to-end.