ONIOKODeveloper

Overview

What ONIOKO offers developers, what it deliberately does not do, and where to start.

ONIOKO builds Extended Human Cognition: technology that perceives observable signals from face, voice and body over time, and keeps what it perceived as memory with its source intact. Products are built on top of it. Partners integrate it.

Every content or conversation tool today reads two things: the text and the numbers. ONIOKO adds a third reading, the person in the video or in the room, and a memory that holds these impressions together over weeks. The human always keeps the interpretation.

The four layers

LayerWhat it doesFor partners
PerceptionReads observable signals from face, voice and body, without labels.Submit or stream consented inputs, receive measured signals or an unavailable state.
MemoryConnects the current context with long-term memory, source and provenance intact.Write and retrieve person-bound records.
CognitionPlaces memory next to current questions, sources stay visible.Used inside ONIOKO products today.
PresenceExtended Avatars with face, voice and continuity across conversations.Used inside ONIOKO products today.

What ONIOKO does not do

  • No emotion labels. Signals are described, never named as feelings.
  • No scoring of people, no ranking, no population scoring.
  • No medical or psychological diagnosis.
  • No continuous recording. Only chosen moments, with consent that can be revoked and leads to deletion.

These are not a legal footnote. They are constraints built into every call: without valid consent for a subject, Perception answers "unavailable" and Memory writes nothing.

Where to start

Honest or absent

This documentation shows only what exists in the code. A resource that is not released appears as planned or migrating, never as a placeholder that looks live. The API reference is generated from the OpenAPI contract of the platform layer; nothing in it is typed by hand.

On this page