Read closely.
Think out loud.

An AI-scaffolded reading tool for academic texts. Highlight what matters, leave notes in the margin, and think through hard passages with a reading coach that asks questions instead of writing your answers.

A research instrument by Jewoong Moon (UA) and Joomi Lee (U Arkansas).

Highlight & annotate

Select a sentence to mark it as important, confusing, or evidence. Write a margin note. Your annotations persist across sessions.

A coach that asks

An AI reading coach probes for evidence, mechanism, and interpretation. It refuses to summarize or write submission-ready text.

Structured reflection

At the end of each reading, your highlights and comments assemble into a draft reflection note. Edit, finalize, submit — it is yours.

Your library

Two short readings to try, paired analytical and narrative. Both work without an account.

Built for research

Reading Companion is also a measurement instrument: every meaningful action emits a timestamped telemetry event so reading behavior can be analyzed alongside reflection quality. The three conditions (system-initiated, user-initiated, unmediated) operationalize a single construct — Locus of Regulation in AI-Mediated Reading.

If you're a researcher or instructor working with us, the views below are for you.