Marketing Engine · plain-English guide
How to read Editor
Editor fact-check & quality gate. This guide explains what it produces, what each part means, and how it stays honest — so you can act on the output with confidence.
Takes
Draft + evidence table
Gives
Edited draft + proof check
The data it uses: The draft and the evidence table.
Watch how it works
How to read the output
Edited draft
The draft after fact-checking — fabricated facts removed, genuine analysis kept.
Fact vs analysis
Verifiable claims that trace to evidence versus brand point of view that is allowed to stand unsourced.
Composite gate
The quality score. Deterministic checks dominate, so a draft cannot pass by satisfying a rubric alone.
The terms it uses
- Composite gate
- A weighted score where grounding, proof, and ban-list checks count for most of it — subjective quality is only a tiebreaker.
- Deterministic-dominated
- Most of the score comes from checks a machine cannot flatter: evidence grounding, proof counts, pattern bans.
- Publish gate
- The one test no model can run: would someone who knows you recognize this as your thinking? Only you can answer it.
How it stays honest
Every agent separates what it measured from your data from what it inferred. You always know which is which.
Measured
- ✓Evidence grounding
- ✓Proof-mix counts
- ✓Ban-list patterns (computed)
Inferred
- ~Edge quality
- ~Voice scoring
- ~Rewrite decisions
Questions people ask
What is the human publish gate for?+
Anti-slop. An LLM can satisfy every rubric and still produce sophisticated slop. The final call — is this your thinking? — stays with you.
Now read the real thing.
Run Editor on your own data and compare the output to this guide. If it matches, you can trust it.