HAVEN, by Justice Apex Inc.
Every child is trying to say something.
HAVEN is a patent-pending swarm-intelligence system that learns what minimally speaking and nonverbal autistic children are trying to say, from their attempted words, sounds and movements, and it's designed to scale to every such child. It's built, and its intelligence is running today on recordings regularly fed to it from its first family's child.
HAVEN starts with ordinary home video and is built to use the shows a child watches as a clue. Each part below says where it stands.
How the intelligence works
Confirmed words
HAVEN finds the sounds a child repeats in ordinary home video and groups them. When a caregiver confirms what a sound means, HAVEN learns from that meaning. Caregiver confirmation is how the confirmed words are built, and it's the method planned for the families in a later trial.
Running now.
Where the words may come from
Some children pick up words and phrases from the shows they watch. HAVEN holds the caption transcripts of a set of online children's videos. It's built to compare the sounds a child makes with what was said on screen, to find the phrase the child may be reaching for. A match is a clue and never proof.
Built and being brought online.
Movement and mimicry
HAVEN's body and gesture tracking is built and has been tested on home video. Matching a child's movements to the shows they imitate is built and being brought online. Mouth shape is the piece still in development. We're working on a way for it to help rule out words a child's mouth couldn't have been forming.
Tracking is built and has been tested on home video. Matching is being brought online. Mouth shape is in development.
Swarm intelligence
These parts don't vote alone. HAVEN's learning loop writes down each part's guess before the answer is known. It then checks each guess against what a person who knows the child confirms, and adjusts how much to trust each part. That is how HAVEN is built to keep learning about a child over time. It's designed to share what it learns across families.
The learning loop is running now. Sharing across families is designed to, and isn't live.
This figure isn't a child, and it isn't made from any recording.
HAVEN's body and gesture tracking is built and has been tested on home video.
A schematic illustration of two identical figures drawn as joints and lines, with no eyes, hair or clothing, and a short line in each round head where a mouth would be. One figure performs a short gesture and the second follows a moment later, beside a panel of simple abstract mouth shapes.
Where HAVEN stands
The next step is a trial with more families. It requires independent ethics review (an IRB) before any other family's data is collected. This funding round goes toward further research, the equipment and requirements for data gathering, and growing the team with engineers and researchers.
What HAVEN isn't
HAVEN supports families. It doesn't diagnose, and it doesn't replace a clinician, therapist or teacher.
Four provisional patent applications were filed June 25, 2026.
Talk with Chad
If you're a funder, researcher or partner and want to talk about HAVEN, write to Chad at chad@justiceapex.ai.
Please don't include details about a child or any health information in your message.