Credits

How Casaluz was made.

Last updated 25 October 2026 · Fisk LLC

Casaluz is made by one person. Some of what is inside it was produced with AI tools, and this page says which parts and which tools, because a parent choosing an app for a small child is entitled to know what they are looking at and listening to.

The letter videos

The 28 alphabet films are generated video, made in September 2026 with Dreamina Seedance 2.5 on BytePlus ModelArk. The child who appears in them is not a real child, and the files carry Content Credentials (C2PA) recording that provenance.

The voice

Every spoken word, phrase, letter sound and lesson line is synthesised speech — Google Cloud's es-US-Neural2-A, a Latin American Spanish neural voice. It is not a human narrator.

What matters for a child is that it is one voice and always the same one: all 3,955 clips are produced once, checked, and shipped inside the app, rather than handed to whatever speech engine each phone happens to have. That is why a word sounds identical on every device, why it works with no connection, and why the pronunciation is consistent rather than drifting between manufacturers.

The words and the teaching

The vocabulary, the phrases, the lessons, the alphabet sequence and the decisions about what is taught in what order are editorial choices, reviewed by a person. The alphabet follows the método silábico used across Latin America — vowels first, then M, P, S, L, T — rather than alphabetical order.

Why say any of this

Because the alternative is letting a parent assume there is a human narrator and a filmed child when there is neither. The tools do not make the app worse; pretending they were not used would.

Questions