Sing to Learn

A family's second language usually dies quietly, one skipped bedtime at a time.

Most bilingual households start with real intention, and lose the second language within a generation. A classroom, or a generic language app, can only meet a child at one level for everyone in the room at once — which means it is wrong for most of the room, most of the time. The one ritual every tired parent already keeps — the bedtime song — gets treated as unteachable, simply because nobody looked at it as a lesson.

The turn

Sing the same lullaby you already sing — the words just arrive in two languages.

Press and hold, and a familiar melody plays while the lyrics cross-fade from the home language into the target language, one line at a time, exactly the way the house already sings it. The microphone listens while it is held, entirely on the device, tracking pitch and pronunciation to know how the singer is actually doing — never a quiz, never uploaded. Melody carrying new words into an old, trusted ritual is the whole bet: a lullaby a parent will repeat hundreds of times without ever feeling like homework.

Teaching at the right level

The Level Map has five named steps: First Hush (the song stays in the home language, one target word planted inside it), Half-Light (one full line arrives in the target language), Rising Moon (the whole chorus turns over), Wide Awake (only the hardest phrase still holds out in the home language), and Sung Through (the whole song, home language dropped). Nobody chooses a level by hand — the app moves a singer up or down the ladder itself from what it heard the night before, a night-by-night re-assessment no classroom of forty could manage for one child. That continuous re-leveling is a direct application of Teaching at the Right Level, an approach developed by the Indian NGO Pratham and evaluated across fifteen years of randomized controlled trials by J-PAL researchers — among the most cost-effective learning interventions ever measured, reaching more than eighty million children. source The mechanism is different — nightly audio instead of a classroom test — but the underlying argument is the one every venture in this portfolio makes the same way. See the full methodology Whether singing itself measurably speeds up second-language acquisition, beyond making the practice something a family actually keeps doing, is a question the company treats as open and worth researching properly — not a settled result to advertise as science.

Where it lives

Sing to Learn doesn't live in an office or a classroom — it lives in the ten minutes before a child's light goes out, in whichever house is singing that night. A typical week: Monday and Tuesday hold at Half-Light while a new line settles in; a rough Wednesday session and the app quietly holds the level rather than pushing forward; by Thursday the chorus turns over cleanly and Rising Moon unlocks; Friday and the weekend repeat it until it's easy, and by Sunday night the Nightly Streak reads seven — the only number either of them was ever chasing. The household, not any one learner, is what actually improves: a parent who once only hummed along is, by the end of most weeks, singing the new line first. That bedroom is only one half of the mechanic. Travelling Bands is designed to ride the same songs and the same level ladder into a school hall in person — the part a phone can't do alone — while this app is what keeps the lesson going once the trailer's pulled away. See Travelling Bands