Reading levels

Stories that read at the right level.

A four-year-old and a fifteen-year-old don't want the same story. Good — they don't get it. Here's how we keep each kid in their own reading world.

Same story, four ways

One plot beat.
Four reading levels.

A 5-year-old and a 14-year-old don’t read the same prose. Tap the tier; the words change. The story doesn’t.

Little

Ages 4–6

The brass spyglass. Not gold. Not shiny rocks. The brass spyglass. It was packed up nice. The words on it matched the map. “Wow,” said Felix. “This is better than gold,” said Mira.

FK < 2.0 · short sentences · familiar words

Kid

Ages 7–10

The brass spyglass. Not gold. Not jewels. The brass spyglass — packed carefully, labeled in the same handwriting as the map. “Huh,” Felix said. “This is better than gold,” Mira said finally.

FK 2.0 – 7.7 · CCSS grades 2–5

Tween

Ages 11–13

The brass spyglass. Not gold coins or gemstones. The brass spyglass — carefully preserved, each item labeled in the same precise handwriting as the map itself. “Huh,” Felix said softly. “This is worth more than gold,” Mira said finally.

FK 6.5 – 10.3 · CCSS grades 6–8

Teen

Ages 14–16

The brass spyglass. Not gold. Not jewels. The brass spyglass — packed with care, labeled in the same hand that had drawn the map. “Huh,” Felix said. They both studied it in silence. “Explain.” And Mira did.

FK 8.3 – 12.1 · CCSS grades 9–10

Names and objects fill in per-story — your kid's next pirate adventure stars whoever the AI picks for them.

How it works

Three steps,
one strict gate.

No vibes. Every passage every kid reads has been scored against a public standard. We can show our work.

①

Anchor

Each tier is anchored to CCSS Appendix A, the federal K-12 text-complexity standard. No proprietary readability vendor in the loop — the bands and formulas are public.

CCSS · public formulas

②

Author

A baseline KID passage gets re-written for Little, Tween, and Teen with grade-tagged exemplar prose in the prompt. Plot beats stay identical; voice and vocabulary adapt.

AI + grade-tagged exemplars

③

Validate

Every passage is scored three ways — Flesch-Kincaid, Dale-Chall, and sentence-length stats — and must pass a 2-of-3 vote against the tier band before it ships.

3-metric vote, every push

Outcome: in-band → ship · 1-of-3 → warn (human review) · 0-of-3 + ≥1.5 grades off → CI break.

From your end

What this feels like.

The methodology lives under the hood. Here's the part you actually touch.

01

You set the level

Each kid's age picks a default tier. Confident reader? Add a +1 grade offset. Two-tap reset if it's wrong.

02

Stories adapt

Same template, same plot, but the voice and vocabulary match the chosen tier. No filler, no dumbing down.

03

Every passage is scored

Three metrics, public formulas, 2-of-3 vote. Misses warn the author; bigger misses break the build.

04

We listen back

How long a kid spent on a passage is captured as signal — never as a gate. A passage that's 2× slower than peers gets flagged for review, not the kid.

The receipts

Don't take our
word for it.

The standards, formulas, and corpora behind every passage. All public, all citable.

  • Standard

    CCSS Appendix A

    NGA Center & CCSSO (2010). Common Core State Standards for ELA, Appendix A: Research Supporting Key Elements of the Standards.

    The federal K-12 text-complexity standard. Defines our four tier bands across multiple readability metrics. No paywall.

    thecorestandards.org
  • Formula

    Flesch-Kincaid Grade Level

    Kincaid, J.P., Fishburne, R.P., Rogers, R.L., & Chissom, B.S. (1975). Derivation of New Readability Formulas. Naval Air Station Memphis Research Branch Report 8-75.

    One of the three metrics in our 2-of-3 vote. Public-domain US government technical report; the formula is pure arithmetic on words/sentences/syllables.

  • Formula

    Dale-Chall Readability

    Chall, J.S., & Dale, E. (1995). Readability Revisited: The New Dale-Chall Readability Formula. Brookline Books, Cambridge MA.

    Second metric. Weighs vocabulary frequency against a ~3,000-word familiar-word list — catches what Flesch-Kincaid misses on short sharp prose.

  • Reading time

    ORF (Oral Reading Fluency) Norms

    Hasbrouck, J., & Tindal, G. (2017). An Update to Compiled ORF Norms (Technical Report No. 1702). Behavioral Research and Teaching, University of Oregon.

    Tier-by-tier words-per-minute benchmarks behind the reading-time signal. We use these for the parent-side outlier surface (CHK-060) — never to gate a kid mid-story.

    brtprojects.org
  • Corpus

    CommonLit grade-tagged library

    CommonLit. Grade-tagged passage library. Used under CC BY-NC 4.0.

    Real published prose at each grade level — the in-context exemplars that shape how the AI sounds at each tier. Excerpts cited inline in the author prompt.

    commonlit.org

Want the full architecture write-up — bands, validator code, failure modes, deferred work? Read ADR 0001 on GitHub →

Try it with your family.

Free during beta. Import a family or start from scratch — takes five minutes. Your fridge will thank you.

Start free →Sign in