KhaiTriOS · Knowledge Graph

Connected knowledge should be inspectable and navigable.

The graph makes relationships visible across principles, projects, playbooks, decisions and lessons without requiring a database or exposing the private memory layer.

Why a graph layer

A library answers “what exists.” A graph also answers “what is related.”

KhaiTriOS already had structured records, but their relationships were distributed across page copy. The graph turns those relationships into explicit public metadata that can be explored by people and reused later by AI systems.

The implementation remains intentionally lightweight: one JSON dataset, one small enhancement script and direct links to canonical records. The graph does not replace the source pages and does not infer hidden relationships automatically.

Interactive explorer

Filter the system, select a node and inspect direct relationships.

The explorer is rendered from the same JSON dataset available as a public artifact.

11 nodes18 edges5 types

Loading public graph data…

KhaiTriOS public knowledge graph Connections across principles, projects, playbooks, Decision Record 001 and Lesson Record 001.

Relationship vocabulary

Edges describe why two records belong together.

GUIDES

A principle or lesson changes future behavior

Used when a record provides a rule or standard for another record.

INFORMED

Project experience shaped a reusable playbook

Used when implementation evidence contributed to a method.

SUPPORTS

A record provides evidence for a decision or system

Used for direct structural or outcome support.

APPLIES

A playbook operationalizes a decision

Used when a method turns a decision into repeatable behavior.

PRODUCED

A decision and its outcomes generated a lesson

Used for the decision → lesson evidence chain.

EVIDENCE

A project record supports a lesson claim

Used when the project is an inspectable evidence source.

Graph data contract

Keep graph metadata small, explicit and source-linked.

NODE

One canonical record

ID, type, title, short label, URL, summary, tags and visual coordinates.

EDGE

One typed connection

Source ID, target ID, relationship type and a human-readable label.

VALIDATION

No broken references

Every edge endpoint must exist and every node URL must resolve to a public record.

OWNERSHIP

Source pages remain authoritative

The graph summarizes relationships; it never replaces record-level evidence or rationale.

Connection examples

The graph now exposes relationships already supported by the records.

PROJECT → PLAYBOOK

Advertising Intelligence informed Growth & Measurement

Dashboard, tracking and pacing problems were translated into a reusable review system.

PROJECT → DECISION

Airbnb Operations supports DR-001

The public record preserves architecture while excluding private apartment and guest data.

DECISION → LESSON

DR-001 produced LSN-001

Later evidence converted the architecture choice into a five-stage publishing rule.

LESSON → PLAYBOOK

LSN-001 guides Automation & AI

Private context must be classified, minimized, validated and human-approved before use.

Phase status

Decision Graph & Memory is complete.

The public KhaiTriOS foundation now includes structured records, evidence-backed lessons and a lightweight relationship graph. The next layer belongs in the protected private vault.

View completed public roadmap