Every business book assumes you have a desk. Mine assumes I have a pocket. BEARTH — five brands, a daily AI briefing and a growing product archive — runs phone-primary, between Basel and Nairobi. The phone this pipeline was built and is documented on is a Samsung S20 FE. This is the actual daily pipeline, not the idealized version.
Why phone-only — isn't that just a handicap?
The constraint is the point. A phone-first operation forces three disciplines that laptop businesses avoid: radical simplicity (if a workflow needs three monitors, it's too complicated to survive), portability (the business crosses borders as easily as I do) and honesty (every tool in the stack has to earn its place on a 6.5-inch screen).
I spent years as a carpenter. On a building site you learn that the craftsman with the fewest, sharpest tools finishes first — not the one with the biggest van. The phone is my sharpest tool. Everything else got cut.
What does the daily pipeline look like?
Production starts at 08:00 CET and the episode publishes at 15:00 CET — active work stays under two hours. The run: deep research scans the last 24–48 hours of AI and remote-work movement and keeps exactly one monetizable signal. That signal becomes a product blueprint, then a scripted 10-minute briefing checked against quality gates, then an openly disclosed HeyGen avatar video, then a $49 product package on Gumroad — SOP, prompts, templates.
The rule that keeps it alive: never skip the day because a secondary asset is late. If the short clips aren't ready by 15:00, the longform and the drop still ship. Daily publishing is the goal; perfection isn't the gate.
Which apps actually carry the operation?
Research runs on Perplexity. Video production is HeyGen (avatar and voice, openly disclosed in every episode) plus CapCut for edit and captions. Publishing happens directly in the YouTube, TikTok and Instagram apps — manual, no posting automation during the proof phase. Commerce is Gumroad: direct product links only. Planning lives in Google Docs and Sheets.
That's the whole stack. No app survives in it that hasn't been used daily for weeks — the same test a chisel gets in the workshop.
What would a laptop actually add?
Honestly: some speed on long documents, a bigger timeline for video edits. What it would cost: the discipline. The moment the operation depends on a desk, it stops being an operation I can run from a matatu in Nairobi or a kitchen table in Basel. The phone-primary constraint is why the system fits into two hours — a laptop would just give the two hours somewhere to leak.
If you want the output of this pipeline instead of the theory, that's exactly what BEARTH SIGNAL is: one signal at a time, turned into a deployable product.