Loading page
About this tool
Grade writing against a defined brand voice on six anchored 1-5 scales and get a weighted consistency score with drift direction.
A tone consistency rubric scores a piece of writing against a defined brand voice on six anchored 1-5 scales — formality, warmth, directness, energy, jargon density and humour — and returns a weighted 0-100 alignment score plus the direction of every miss. It uses a behaviourally anchored rating scale, so each level carries a concrete description instead of a vague label, and each dimension is normalised by the worst possible miss for its own target. Useful for content leads, editors and anyone reviewing AI-generated copy at volume.
Open Tone Consistency Rubric on AltFTool — it loads instantly in your browser.
Paste or type the text you're working with.
Choose how it should be transformed or analyzed.
Copy the finished text into your document or post.
Every 1-5 level has a written description, so scores are repeatable across raters.
Set weights per dimension; a dimension weighted zero drops out of the score entirely.
Reports which dimension costs the most points and in which direction it drifted.
Define a target level on each voice dimension, rate the sample on the same scale, then score alignment as the weighted average of one minus the miss divided by the worst possible miss for that target. A target of 3 can only be missed by 2 points while a target of 1 can be missed by 4, so normalising per dimension stops extreme targets from dominating the score.
The six used here cover most published voice guides: formality, warmth, directness, energy, jargon density and humour. Nielsen Norman Group's tone-of-voice work uses four similar dimensions (funny to serious, formal to casual, respectful to irreverent, enthusiastic to matter-of-fact); the extra scales here separate jargon and directness because those are where AI-generated copy drifts first.
90 and above reads as on voice — the remaining difference is inside normal rater noise. 75 to 89 is minor drift fixable with a couple of edits, 60 to 74 is drift a regular reader would feel, and below 60 the piece effectively reads as a different brand and is faster to rewrite than to patch.
Have both score the same sample and compare exact agreement and adjacent agreement (within one scale point). Adjacent agreement below roughly 80% almost always means an anchor is ambiguous rather than that a rater is wrong — rewrite the anchor for the dimension with the biggest gap and re-test.