Core platform

What Heirloom does

Five integrated modules around the KAIROS regulation engine. Every one of them reads the same load vector, so the app never asks more of you than you have to give it.

Cognitive engine

KAIROS

The intelligence underneath Heirloom. KAIROS decomposes mental effort into its intrinsic, extraneous and germane components, then adjusts the whole experience against that vector. It carries the Focus Pipeline for signal filtering, ALETHEIA for epistemic protection, and CORTEX for allostatic stress monitoring.

Read the Kairos page

Active regulation

Regulate

Load management with feedback you can act on: focus recommendations, autism-adapted regulation strategies, and a reading that spikes before you notice you are spiking.

AI companion

Companion

A conversational partner that adapts tone, pacing and complexity to your current state. Not a chatbot — a context-aware support system on GPT-4, with voice through AWS Transcribe and Polly.

Digital wellness

Wellness

Mood tracking, journaling and goals that read neurodivergent patterns rather than flattening them. Insights surface the connections between load, emotional state and daily rhythm.

Safety architecture

Graceful degradation

Error boundaries that degrade rather than disappear. Every component is built on the understanding that for this user, reliability is a safety property, not a polish item.

Method

Cognitive Load Theory, applied

KAIROS splits mental load into three measurable channels and uses that vector to drive every decision in the app.

01 Task difficulty How inherently complex is the thing in front of you? KAIROS measures it in context and moves its recommendations against that reading. intrinsic

02 Environmental noise The overhead your surroundings, a badly built interface or sensory input adds on top of the task. This is the channel Heirloom actively works to bring down. extraneous

03 Useful learning effort The productive work that builds understanding. KAIROS protects this channel and spends the other two down to make room for it. germane

Cognitive Load Theory is Sweller's, not ours. What is ours is the decision to treat the three channels as a live signal a running application can read, rather than a framework for designing a lesson once and walking away.

Under the hood

Built to last

Production-grade infrastructure with safety as a first-class concern. Every layer is designed for the people who depend on it.

Stack

React Native / Expo

Cross-platform client, one codebase

KAIROS (FastAPI)

Real-time cognitive load filtering

Node.js / Express + MongoDB

Structured logging, in-memory fallback on database outage

Supabase (LAZRUS) / Firebase Auth

Identity and shared platform data

GPT-4 · AWS Transcribe / Polly

Companion reasoning and voice interaction

W33KND

Telemetry and clinical oversight console

kairos/load.py

# The load vector is three numbers, not a mood.
class LoadVector:
    intrinsic:  float   # task difficulty
    extraneous: float   # environmental noise
    germane:    float   # useful effort

def budget(v: LoadVector) -> Surface:
    headroom = CAPACITY - (v.intrinsic + v.extraneous)
    if headroom < MARGIN:
        # Spend the noise channel first. Never the germane one.
        return Surface.reduced(strip="extraneous")
    return Surface.full(protect="germane")

Heirloom is a wellness tool. It does not diagnose, it does not treat, and it is not a substitute for care from a clinician.

Your brain works differently. Your tools should too.

Heirloom is in development. Write to us and we will put you on the early-access list.

Request early access Heirloom on PRJCT LAZRUS

Cognitive regulation, designed for autism.