RepStack

Editorial standards

How these pages get made.

This site publishes slowly and explains itself. Below is the bar every page has to clear, what is out of scope, how AI is and is not used, and what happens when something is wrong.

Last updated June 10, 2026

The bar

Six things every guide must have.

  1. 1

    One decision per page

    Each guide answers one question a lifter actually faces. No ultimate guides, no topic sprawl.

  2. 2

    Something only this site has

    The app's actual progression rules, an original diagram, a real screenshot, or patterns from training logs. A page with nothing proprietary does not ship.

  3. 3

    A worked logbook example

    Real numbers in real notation — 80 kg x 8, 8, 7 — not abstract advice.

  4. 4

    Sources next to the claims

    Numbers, comparisons, and evidence claims are cited inline where they appear, plus a sources list at the bottom. Three to six sources for a normal guide.

  5. 5

    A 'when it fails' section

    Every rule has edge cases. The page names them instead of pretending the advice is universal.

  6. 6

    Visible author and dates

    A real name, a published date, an updated date, and a next-review date — on the page, not hidden in code.

Scope

Training guidance for healthy adults. Nothing else.

The guides cover workout logging, progression, programming decisions, and how RepStack's own engine works. They do not cover injuries, rehab, pain, medical conditions, pregnancy, training for minors, or supplements. Those topics need qualified professionals, and this site does not pretend to be one. If a future page touches a topic that needs expert review, the reviewer is named on the page — no review, no page.

Process

Brief, sources, draft, then a cold read.

Every guide starts as a brief: the decision it answers, the claims it will make, and the sources that back them. The draft is written against that source package, then fact-checked claim by claim on a later day before it publishes. Training-science pages are reviewed at least annually; pages about RepStack's own methodology are re-checked whenever the app's rules change. Each page shows its next scheduled review date.

AI policy

AI helps with research. It does not get a byline.

AI tools are used the way a research assistant would be: collecting sources, stress-testing outlines, and arguing against drafts. The published voice, the examples, the judgment calls, and every claim are the author's, verified against the cited sources. No page on this site is auto-generated, and no page exists because a keyword does.

Disclosures

RepStack is our product. The pages say so.

This site exists to support the RepStack app, and pages that show the app say so plainly instead of staging neutrality. There are no affiliate links and no sponsored posts; if that ever changes, the disclosure appears before the link, not in a footer.

Corrections

Wrong gets fixed, visibly.

Spotted an error? Email alexandrutirim@gmail.com. Anything safety-relevant is triaged the same day. Factual errors are corrected within 48 hours and the page carries a visible note saying what changed. Ordinary fixes — typos, broken links — happen within a few days without ceremony.

The standards themselves live here, in public, so you can hold the guides to them. Who writes all this is on the about page.