Skip to main content
ReadLens
Open demo

CHI 2027 submission

In-place contextual assistance for second-language reading

ReadLens combines contextual translation, prewritten questions and annotations for selected words and passages.

Open demo PDF

Approach

Four interactions around the selected text.

  1. Select

    A selected phrase is highlighted within the article.

    Select a word, expression or passage.

  2. Translate

    A brief contextual translation appears beside the selection.

    Read a brief translation in context.

  3. Explain

    A prewritten question opens a contextual explanation.

    Choose a question or write your own.

  4. Save

    A saved response is attached to its marked source text.

    Attach a response to the selected text.

Interface

Select text, request explanations, save annotations.

Complete research interface figure: reading and contextual translation, a further-assistance panel with prewritten and generated questions, and annotation review, with a legend identifying the controls.
Study interface; labels and example responses shown in English.

The public demo uses French articles with English assistance; the study used English articles with Italian assistance.

Open demo

Comparison

The baseline combines dictionary lookup and article-context chat.

Bilingual dictionary

The baseline opens a bilingual dictionary entry in a popup beside the selected word.

Double-click a word to open a bilingual entry.

Contextual chat

The baseline shows the article on the left and a separate chat panel on the right.

Reader-written questions, with the article as context.

Study

A within-subject, counterbalanced comparison.

The design drew on 89 survey responses and interviews with 10 readers and 3 language experts.

Eighteen Italian-speaking readers followed one of four routes across two readings. Each route pairs different English articles A and B with ReadLens and the dictionary-plus-chat baseline, crossing article and interface order.

Requested explanations

10 of 18 readers requested a wider range of explanations with ReadLens.

Explanation functions requested in each condition, shown as shares of questions and numbers of readers.
Requested explanation functions, by condition.
Request counts and ranges of explanation functions for each of the eighteen readers, comparing ReadLens with the baseline.
Request counts and function repertoires for all 18 readers.

Two readers contributed 31 of the 59 ReadLens questions. These counts do not establish improved comprehension.

Interview excerpts, translated from Italian

Requesting assistance

“without the stress of having to sit there and write the right question, make the right request.”

P02

Selective use

“I used only the translation and did not use the other features because, at that moment, I did not need them.”

P16

Citation

BibTeX

@misc{readlens2027,
  title = {ReadLens: In-Place Contextual Assistance
           for Second-Language Reading},
  year  = {2026},
  note  = {Manuscript under review at CHI 2027}
}