Communication support that grows with the person

More freedom
to communicate.
A more connected life.

KIN provides private, real-time communication support for persons with Autism and other neurodiverse adults—helping strengthen timing, clarity, relevance, and reciprocity while keeping every choice with the user.

Join as a tester → ▶ See KIN in action

Autonomy first NLP + LLM intelligence No raw conversation retained

A young adult using a wired earbud while bagging groceries during an everyday conversation

KIN is supporting Following the conversation

Private cue

“You could ask: how did you sleep?”

You decide what to use.

Family conversations• Grocery stores• Work• Appointments• Friends• Community life• Family conversations• Grocery stores• Work• Appointments•

Built for life as it happens

One private layer of support.
Many kinds of moments.

KIN is being built for ordinary conversations across home, work, relationships, errands, and community life—not only for practice sessions.

At the store

Choices, questions, checkout, and the small exchanges that make community life possible.

With family

Everyday small talk, practical questions, shared stories, and familiar routines.

At work

Meetings, task clarification, coworker check-ins, and professional collaboration.

On dates

Getting to know someone, handling pauses, and staying present without scripted performance.

With friends

Shared interests, humor, emotional disclosures, and meaningful back-and-forth.

In the community

Appointments, services, transportation, classes, and daily independent living.

Our mission

Help persons with Autism communicate, connect, and thrive—on their own terms.

KIN does not speak for the user. It creates a private layer of support when a conversation becomes difficult: a pause runs long, an answer becomes tangled, a topic drifts, or a natural opportunity for reciprocity appears.

Support, not substitution. Guidance, not control.

Everyday life

Communication does not happen in a clinic.

KIN is being designed for the ordinary moments that shape independence, belonging, and opportunity.

A neurodivergent adult using a wired earbud while bagging groceries

01 · Everyday errands

Bagging groceries. Asking for help. Joining the moment.

The first context is intentionally ordinary—and deeply meaningful.

An adult wearing a small earpiece while speaking with a family member at home

02 · At home

Family small talk

Support during the conversations that happen every day.

A person speaking beside a futuristic communication interface

03 · Beyond home

Work, community, relationships

A platform designed to expand with the user’s goals.

The KIN experience

Present in the moment.
Private by design.

KIN begins when the wearer chooses. It listens only during an active session, offers optional support, and returns progress to the wearer—not to a scorecard of “normal” conversation.

One-tap control Optional private cues Personal growth insights

KIN is ready to activate a private communication coach

KIN is following the conversation during an active private coaching session

KIN displays personal communication growth insights

KIN shows a completed private coaching session summary

Live, only when invited

How KIN works

Understands the moment.
Learns the person.

KIN combines natural-language processing, large language models, and a real-time conversation-state engine. The goal is not to force one “correct” style—it is to learn when, where, and how support is useful for this person.

A visual representation of speech becoming context-aware private audio guidance

Live conversation state

Turn state84

Intent72

User fit91

Individual memoryConcrete examples work best in choice questions

Next best actionOffer one optional, familiar phrase

01
You activate support
A single press turns the current prototype on or off. The user decides when KIN is present.
02
Speech is processed in the moment
KIN follows timing, turns, repetition, relevance, and the communicative request without retaining the raw conversation by default.
03
NLP and LLMs interpret context
Natural-language processing and large language models help distinguish a direct question, a tangent, a completed answer, confusion, or a reciprocity opportunity.
04
KIN learns the user’s idiosyncrasies
With consented use and validated outcomes, KIN is designed to learn common expressions, pacing, preferred cue language, and the contexts where help is—or is not—useful.
05
One private option reaches the ear
A warm voice offers a brief example or action. The user can use it, change it, or ignore it.
06
Human feedback keeps learning grounded
Adaptation should be evidence-based, auditable, reversible, and guided by each person with Autism’s own experience.

01Live speechPermission-based

02Conversation stateTiming + context

03NLP + LLMsMeaning + intent

04User memoryWhat helps this person

05Private cueUser remains in control

Privacy by architecture

Your conversation is processed—not kept.

The current beta does not retain continuous raw audio or complete conversation transcripts by default. KIN stores only pseudonymous structured performance signals and generated cue text needed to evaluate reliability. Any future research recording requires separate, explicit consent.

✓No continuous audio storage ✓No full transcript storage ✓Research recording is opt-in

Domains of communication KIN supports

Specific intelligence for specific moments.

Communication is not one skill. KIN considers several domains together before deciding whether support is useful.

Timing

Long pauses, run-on speech, speaking pace, turn entry, and turn completion.

Repetition

Recognizing repeated words or ideas and helping organize the main point.

Direct answers

Helping answer the communicative request without unnecessary drift or delay.

Relevance

Noticing tangents and offering a gentle path back to the current topic.

Clarification

Providing language to ask for repetition, explanation, or more time.

Reciprocity

Supporting acknowledgments, related follow-ups, and natural conversational back-and-forth.

Expanding domains

Repair · topic initiation · topic maintenance · topic shifting · boundaries · emotional regulation · cue burden

KIN combines signals across domains. A four-second answer may be entirely appropriate; a three-second answer may already be repetitive or off-topic. Time begins an evaluation—it does not automatically trigger a correction.

A context-aware coach

Not “ask a question.” Something useful here, now.

Explore simplified examples of how the system thinks about different moments.

KIN notices repetition and a practical choice is still unresolved.

Private audio suggestion

“A good example is: I’d like cereal.”

Optional. Adaptable. Never spoken to the other person.

Human validation first

Human validation is not a footnote. It is the standard.

Human validation is a major focus of KIN. The software can surface a pattern, but each person with Autism’s experience determines whether that pattern was meaningful, whether the cue helped, and whether the system should adapt.

User authorityThe user remains the primary source of truth.

Auditable learningChanges should have evidence and be reversible.

Consent by designResearch recording is separate and explicit.

A person speaking while a visual system maps communication signals

Session review2:14 PM

KIN detected a reciprocity opportunity after a family check-in.

Illustrative future validation interface

Concept render of a future discreet KIN earpiece and clothing clip

Hardware built around the person

Today: an iPhone and wired earbuds.
Tomorrow: KIN hardware.

The first pilot intentionally uses hardware people can access now. The longer-term vision is a discreet KIN earpiece, lavalier-style clip, or coordinated system designed with persons with Autism—not merely for them.

One-press agency Private audio Sensory-aware design Visible listening state

Concept imagery. Final hardware design has not yet been manufactured.

Our story

A human-centered idea—
years ahead of the models.

KIN is the continuation of a career spent at the intersection of neurodiversity, human behavior, acute medicine, and technology.

KIN by Enti3 Founder Story Christopher Taicher, MD Former behavioral-therapy professional · Harvard-trained emergency physician · Technologist

Early behavioral-therapy experience

It began with direct support.

Before medical training, Christopher Taicher worked in behavioral therapy supporting persons with Autism—years before becoming a physician or technology founder.

Emergency medicine

Then came medicine at the highest level.

He became an emergency medicine physician and completed the Harvard-affiliated residency at Massachusetts General Hospital and Brigham and Women’s Hospital.

Technology architecture

The system was conceived before today’s LLMs.

Before modern language models made real-time coaching practical, he designed Enti3’s original natural-language-processing architecture for contextual communication support.

KIN by Enti3

Now the technology can meet the vision.

KIN combines live conversation signals, NLP, LLMs, private audio, structured outcomes, and human validation to support more self-determined communication.

Early recognition

The concept drew institutional and media attention while the enabling technology was still emerging.

PROCOPIO

2022–2023 LaunchPad

Selected by Procopio

Enti3 was chosen for Procopio’s LaunchPad incubator, where emerging companies receive corporate, intellectual-property, and growth support.

Read the LaunchPad announcement ↗

SAN DIEGOBUSINESS JOURNAL

November 2022

Featured by the San Diego Business Journal

The work and its founder were recognized in the publication’s 2022 Technology Achievement coverage.

View San Diego Business Journal coverage ↗

Help shape what comes next

We’re inviting a small group to test KIN.

We are looking for persons with Autism and other neurodiverse adults—and a family member willing to have ordinary conversations—to help us understand what is useful, mistimed, distracting, or genuinely empowering.

An iPhoneCompatible with the current TestFlight prototype.

Wired earbuds with a microphoneThe cord button can activate or stop support.

A participating family memberFor ordinary, low-pressure conversation testing.

Spoken-language fit for this early prototype Approximately an age-8+ neurotypical spoken-language level. This is a technical compatibility criterion—not a measure of intelligence or potential.

Questions

A quiet technology with clear boundaries.

Does KIN speak for the user?

No. KIN offers one private option. The user chooses whether to use it, change it, or ignore it.

Is KIN medical treatment?

No. KIN is an experimental communication-support prototype and is not an emergency, diagnostic, or treatment service.

Does KIN record every conversation?

No. Continuous raw audio and complete conversation transcripts are not retained by default. The beta may store pseudonymous structured performance events and generated cue text so the team can evaluate reliability and improvement.

Why wired earbuds?

They are inexpensive, private, do not need charging, and provide a physical cord button that can give the user direct control over when support is active.

How will KIN become individualized?

With consented use, KIN is designed to learn the person’s common expressions, pacing, preferred wording, and the contexts where support is useful—or unnecessary. Adaptations should remain bounded, auditable, reversible, and human-validated.

KIN by Enti3