What is the weekly loop in a Content OS?

It's one recurring review that runs the three apps together: the tracker surfaces this week's niche outliers, the vault hands you the exact hooks and CTAs that made them hit, and your dashboard shows which of those patterns actually work for you. The output is a dashboard plus three content pillars for the coming week, each with three ideas derived from proven outliers.

How often should I run the content review?

Weekly. That cadence matches how fast the algorithm and your niche move — frequent enough to catch new outliers while they're still fresh, spaced enough that you have real new data each time. Run it the same day each week (Monday planning works well) so it becomes a ritual, not a decision.

Does the system decide what to post for me?

No. The loop suggests — it surfaces outliers, extracts mechanics, and proposes three pillars — but you decide. That human-in-the-loop gate is deliberate. The system is a research and analysis engine that hands you a ranked, evidence-backed shortlist; the creative call and the publish button stay with you.

The Weekly Loop

Three apps sitting in three folders aren't a system yet. The loop is what fuses them. Run one review, once a week, and the tracker, the vault, and the dashboard stop being separate tools and start feeding each other.

What the loop actually does

Each app answers one question. The loop chains those answers into a plan:

The tracker surfaces this week's outliers in your niche — what broke through.
The vault hands you the exact hooks and CTAs that made those posts hit — the mechanics.
Your dashboard shows which of those patterns actually work for you — your numbers as the filter.

→ Out comes three pillars for the week, each backed by real outliers, not vibes.

The prompt

This is the one prompt that runs the whole loop. Paste it into Claude Code once a week:

Run my weekly content-OS review. From track.db + data/mine: (1) my reels this week vs
last week (median views, avg watch time, saves/reach, shares/reach). (2) The top 10
competitor outliers of the week (multiplier >= 3) with account + hook + why it worked.
(3) Cross-reference: which of those patterns I haven't tried yet. Output: an HTML
dashboard + 3 pillars for next week, each with 3 ideas derived from this week's proven
outliers.

Read what it's doing, piece by piece:

Part 1 compares your own week to last — is watch time up? Are saves/reach climbing? That's your dashboard talking.
Part 2 pulls the niche's real winners (multiplier ≥ 3, so only genuine outliers) with the spoken hook and a reason it worked. That's the tracker plus the vault.
Part 3 is the leverage move: it cross-references the winning patterns against what you've already done and tells you what you haven't tried yet. That's where next week's ideas come from.
The output is a dashboard you can look at plus three pillars, three ideas each — a shortlist grounded in evidence.

The human-in-the-loop gate

The loop hands you a ranked, evidence-backed shortlist. It does not post anything. That gate is deliberate.

Warning:

The system suggests, you decide. Nothing auto-posts. The loop surfaces outliers and proposes pillars, but the creative call — and the publish button — stay with you. Every serious builder ships this gate on purpose: the machine does the research, you do the judgment.

Why keep a human in the loop? Because the data tells you what worked, not what's you. A hook can be a niche outlier and still be wrong for your voice. The loop's job is to make sure you're never staring at a blank calendar — you're choosing from a shortlist of proven mechanics. That's a much better problem to have.

Make it a ritual

Pick one day a week and run it every time — Monday planning is a natural fit. The point of a fixed cadence is to remove the decision. You don't ask "should I plan content this week?" You run the prompt, you get three pillars, you pick. The system does the heavy lifting so the ritual stays cheap enough to actually keep.

Knowledge check

What does the weekly loop produce as its output?

Key takeaway

One weekly prompt closes the loop: the tracker surfaces outliers, the vault hands you the mechanics, the dashboard filters by what works for you, and out come three pillars for the week. The system suggests; you decide. Run it the same day each week and it becomes a ritual, not a decision.