Creative intelligence
Once
A grandmother in Alexandria opens a book. In Riyadh, Dubai and London, three children see the same page — and the words light up as she says them.
Not a video call. Not an audiobook.
Video calls put a face on a screen. They don't give a family anything to do together.
Work, study and migration have spread families across cities and continents, and a grandparent watching a pre-reader lose interest in four minutes is not a solved problem. Once ships the storyteller: she turns the page and the child's page turns with it, and the engine follows her real voice through the hesitations, the rereads and the asides — treating the book as a scaffold, not a script.
Live, low-latency voice
Self-hosted real-time audio, with permissions set so only the reader can speak and up to a dozen people in a room. One link for the grandparent — nothing to install, no account to make first.
Alignment we own
Speech recognition runs on our own servers rather than a third-party cloud, and a purpose-built aligner matches voice to text — diacritics included. No external aligner sits in the critical path.
A studio behind the shelf
Publishers build a title end to end — story bible, illustration, layout, narration, ambient sound, continuity audit — then translate the finished book into the other language in a single action.
What a session is
The whole product is one page turning at the right moment.
After the story
The parent gets a note, not a dashboard.
Every session ends with a short written recap in the book’s own language: how it went, three questions to ask at breakfast, and the words the child stumbled over — with the reason. It is written to a privacy rule that forbids quoting a child at length or repeating anything that looks like a name, a school or an address.
Voice studio
Bedtime, in a voice they love — with the consent recorded.
Record a loved one reading for about a minute and the family can hear any book in their voice. The consent is not a checkbox buried in terms: the named speaker is captured, the confirmation is stored with the sample, and revoking it deletes the clone, every sample and the consent recording together. A voice made for one family can never be used in a publisher’s production pipeline.
Arabic-first, by architecture
Arabic is the default. English is the port.
Children's digital reading is overwhelmingly English-first, and rich Arabic read-along barely exists. Once was built the other way round: the site opens in Arabic, the interface mirrors properly rather than being flipped, and the aligner was taught to handle tashkeel before it was taught anything else. Word-level highlighting that survives diacritics was the hardest problem in the product. It was solved first.
مرّة
حكاية واحدة، صوتٌ يضيء الكلمات
القراءة معاً، رغم المسافات
An act of love, rendered as light.
Once Studio
Every idea can become a story.
The reading experience needs books to read. Studio is the other half of Once: a production line that takes a publisher, an educator or a brand from an initial idea to a finished, illustrated, narrated title — then translates it into the other language in a single action.
Story bible
Characters, world and tone, fixed before a single page is drafted.
Pages & cover
Draft, refine and lay out each page against the bible.
Illustration
Art directed per page, with a continuity check across the book.
Narration & sound
Recorded or generated narration, plus ambient sound cues.
Review & publish
A moderation pass and a human approval before it reaches a shelf.
Who it is for: publishers and children's-content creators, educators and students, cultural and entertainment organisations, and brands that want a story rather than an advert.
Reading together, across distance.
Once runs at once.axera-sa.cloud. If you publish children’s books in Arabic, the studio and the shelf are both open to talk about.