Reading Leveler
Adapt any classroom reading in under a minute — one passage, three reading levels, with the concepts kept intact.
- Next.js
- TypeScript
- Supabase
- Clerk
- Claude API
A fifth-grade teacher may be running one lesson across students reading at a second-grade level, on-level readers, advanced readers, English language learners, and students with accommodations. The usual answer is to rewrite the science reading by hand, three times. Reading Leveler does that in about a minute.
Just paste in a passage, and Reading Leveler produces three versions — Support (grades 2–3), Core (grades 4–6), and Extension (grades 7–8) — with inline change markup so you can see exactly what moved, per-paragraph refinement when one paragraph lands wrong, and printed handout support. It’s live at reading.bulrushlabs.com.
Adaptation rather than simplification
“Rewrite this more simply” is a solved prompt. What teachers actually need is an accessible reading where the educational content survives. The student still has to get the key concepts, causality, chronology, factual and scientific accuracy, and the academic vocabulary the unit is teaching. A Support version of the photosynthesis reading that no longer contains the word chlorophyll has removed the lesson.
Reading Leveler has specific support for protected terms. The AI is told which vocabulary must survive the rewrite, and the readability scorer excludes those terms and proper nouns. Otherwise “Declaration of Independence” alone drags the grade estimate up and the tool starts stripping out the very words the reading exists to teach.
Hitting a grade level is an engineering problem
Flesch–Kincaid is 0.39·ASL + 11.8·ASW − 15.59 — average sentence length and
average syllables per word. Protected vocabulary and named entities push
syllables-per-word to roughly 1.35 in real classroom passages rather than the
1.28 you’d assume, which changes the sentence lengths you have to aim for. The
prompts are tuned for that: Extension
needs about 18-word sentences to reliably reach grade 7.5, while 15-word
sentences only get to 5.5. Each level also carries a max-sentence cap, otherwise
a single 30-word sentence could hide behind a good average, and it is
precisely the sentence a struggling reader stalls on.
When a level lands outside its band the app retries with the miss fed back in — but only when the miss is larger than half a grade. That threshold came out of the eval harness: asked to make a small adjustment from a near miss, Anthropic Sonnet 4.6 reliably overshoots by two or more grades. A retry that makes the result worse is worse than no retry, so near misses are left alone and surfaced to the teacher instead.
Prompt changes are evaluated offline against a fixed corpus of science and history passages before they reach production.
How can I use it?
Reading Leveler is in private beta, but the demo is open and doesn’t require a login. If it looks useful for your classroom and you want an account for saving passages and trying the full workflow, send me a note. I’m especially interested in feedback from teachers using it on real material.