ClearSpec

Reviewed 18 August 2026 · next review 17 February 2027

Readability checker for technical documentation

Paste a procedure and see the sentences a reader will stumble over — not a grade level, but the exact spans where the actor is missing, the reference is ambiguous or the step does too much.

Last updated · Written and reviewed by Jonatan Tensetti · Tensetti Tools

Method 1.0.0 · reviewed 18 August 2026 · next review 17 February 2027

Step rules and action clarity are weighted normally; multi-action steps matter most.

0 words

Free text is analysed in this browser and is not sent to our server. Do not paste confidential material on a device you do not control.

What reading-grade scores miss

Classic readability formulas were designed for continuous prose and estimate difficulty from sentence length and word length. Documentation fails differently. A ten-word step can be unusable — "Reset it appropriately" is short, easy to pronounce, and impossible to follow. A thirty-word step can be perfectly clear when it names the actor, the condition and the observable result.

That is why this checker reports 22 specific signals grouped into five categories rather than a single grade. Each finding points at an exact span, states the reason, gives a neutral edit instruction, and where relevant carries a false-positive note so you can dismiss it with confidence.

The five things worth measuring in documentation

  • Actionability. Does each step name who acts, do one thing, and end in a result someone can observe?
  • Ambiguity. Pronouns without antecedents, vague quantifiers, and conditions that arrive after the instruction.
  • Consistency. One term per concept, one pattern per sentence type, and abbreviations expanded on first use.
  • Structure. Sentence length, nesting depth, and steps that hide a second procedure inside them.
  • Safety language. Whether a warning carries the hazard, the consequence and the avoidance in an order readers can scan.

How to use the result

Work top-down through the findings and stop after ten. Fix the actor and ambiguity items first: they change comprehension most per edit. Mark false positives as reviewed so they do not distract on the next pass. Re-run and compare — the same text with the same settings and the same method version always produces the same result, so a change in the indicator reflects a change you made, not a change in the tool.

Privacy, in one line

The free analysis is local. Nothing you paste leaves your browser, and there is no account to create before you try it.

Questions

Why not just use a Flesch reading-ease score?
Reading-ease formulas count syllables and sentence length. A procedure can score well and still be unusable because the actor is missing, the condition comes after the action, or one component has three names. Those are the failures that generate support tickets.
Is anything I paste sent to a server?
No. Analysis runs in a Web Worker in your browser. Source text, highlighted spans and findings are not sent to our server, to analytics or to logs.
How long can the text be?
The free check accepts from 20 up to 1,000 words. Larger documents, full findings and local file opening are part of the paid workspace.
Does a high score mean the document is correct?
No. The checker measures clarity signals only. It does not establish technical accuracy, legal sufficiency or safety, and it never labels content unsafe.

ClearSpec uses an independent proprietary clarity method. It is not affiliated with ASD/STEMG, is not an ASD-STE100 checker or certification service, and does not certify ISO 24495-1 conformance.

References

Checked 18 August 2026. Referenced for context only; no standard text or dictionary content is reproduced.

ClearSpec reports automated heuristics from method version 1.0.0. The clarity indicator is not a measure of technical accuracy, usability, legal sufficiency, safety or standards conformance. A competent person must decide the final wording, especially for safety content.