She knew the sounds. The sentences kept cheating.

Description and use

The readable‑for‑her tool generates sentences that a child can read using only known letter‑sounds, checks existing sentences for readability, and ranks them by sense with a local language model.

Problem

A child learning to read struggles to blend sounds into words, making many beginner sentences unreadable because they contain digraphs or sounds not yet taught.

Outcome

The tool produces readable sentences that match the child’s current phonics knowledge, checks existing sentences for readability, and ranks them for sense, enabling a structured, step‑by‑step phonics progression without external dependencies.

Workflow

• 1. Specify the letter‑sounds the child knows and any sight words.
• 2. Run "readable --stage N --count M" to generate sentences for that stage.
• 3. Use "readable --check" to verify if a given sentence is readable.
• 4. Use "readable --step story" to generate a short story of readable sentences.
• 5. Review the output and decide when to advance the child to the next stage.

Evidence

A child learning to read is taught letter‑sounds a few at a time. A sentence is only real practice if every word in it can be sounded out with the sounds she has so far.

Source

So I built readable-for-her: you tell it which sounds she knows, and it gives you sentences she can read every word of.

Source

The tool has two parts, and the split between them is the whole design. Rules decide what she can read. A word is readable only if it can be cut, left to right, into letter‑sounds she has been taught.

Source

A model decides what makes sense. Rules are happy to write A mat can dig. They have no idea a pig is the one that digs. That judgement is what a language model is for, so each candidate sentence is scored by SmolLM2-135M.

Source