Readability rewriter
Matching text to its audience’s reading level is one of the highest-leverage edits you can make — too complex and consumers bounce, too simple and professionals lose trust. This tool uses your own API key to rewrite any passage to a target Flesch-Kincaid grade level and shows the measured grade before and after so you can confirm the rewrite actually landed.
What the Flesch-Kincaid grade scale means in practice
The Flesch-Kincaid Grade Level converts two sentence-level measurements — average sentence length and average syllables per word — into a number that roughly maps to a US school grade. Here is what each band feels like as readable text:
| Grade | Audience | Typical example |
|---|---|---|
| 4–6 | General public, plain language | Government plain-language communications, NHS patient leaflets |
| 7–9 | Educated adult general audience | Quality newspaper features, popular science |
| 10–12 | Professional / educated adult | Business reports, legal summaries for professionals |
| 13–16 | University-educated specialist | Academic papers, technical documentation |
| 17+ | Specialist expert | Dense academic or legal prose |
LLM output frequently lands in the 13–16 range by default because models are trained on a corpus heavy with academic and technical writing. Rewriting to grade 8 makes the same information accessible to a much wider audience without losing accuracy.
How it works
You paste your own OpenAI or Anthropic key, the text, and a target grade. The
tool sends a single instruction asking the model to rewrite the text to that
reading level while preserving meaning and key facts, calling the provider
directly from your browser (Anthropic requests carry the
anthropic-dangerous-direct-browser-access header). It then computes the
Flesch-Kincaid grade of both the original and the rewrite locally using the
standard formula — 0.39 × (words/sentences) + 11.8 × (syllables/words) − 15.59
— with a heuristic syllable counter, so you get an objective before/after rather
than just trusting the model’s claim.
Illustrative worked example
Take a passage at grade 14:
“The implementation of machine learning inference pipelines necessitates meticulous attention to the parameterization of batch processing configurations in order to ensure optimal throughput characteristics.”
A rewrite targeting grade 7 might produce:
“When you set up a machine learning pipeline, batch size settings matter a lot for speed. Tuning them carefully keeps the system running fast.”
Both sentences convey the same practical point. The rewrite shortens sentences, replaces multisyllabic technical vocabulary with common words, and cuts redundant phrases. The Flesch-Kincaid formula reflects all three changes immediately.
Tips and notes
- Grade 6–8 for consumer copy. Most general-audience web content targets this band; technical documentation often sits at grade 10–12.
- Verify, do not assume. Models sometimes overshoot or undershoot the target, which is exactly why the measured scores are shown — rerun if it missed.
- Shorter sentences move the needle most. The formula is dominated by sentence length, so splitting long sentences is the fastest way to drop a grade.
- Your key never leaves the browser. It is used only for the direct provider request and is never stored.