Dual audience architecture — humans and AI from the same data

LOGOS must serve humans through the UI and AI systems through
the API with equal fidelity from a single source of truth.
No separate data stores, no sync problems, no second-class
access for either audience.

Why It's Needed

Most knowledge systems are built for humans and have an API
bolted on as an afterthought. The API is a degraded view —
it returns less, it lags behind, it requires workarounds to
get full fidelity. When AI systems consume that API they are
working from an incomplete picture.

LOGOS exists simultaneously as a human knowledge system and
an AI training and context source. CORTEX feeds from LOGOS.
AI-assisted implementation reads specs from LOGOS. The chat
import pipeline writes to LOGOS. If the AI-facing interface
is a second-class citizen, all of those capabilities are
compromised.

The Requirement

Every piece of data in LOGOS is accessible through both the
UI and the API with identical fidelity. This means:

• Same data — no fields available in UI that are
  absent from API, and vice versa
• Same timeliness — API reflects current state with
  no lag behind the UI
• Same metadata — all post metadata, edit history,
  state transitions, and cross-references are API
  accessible
• Same event log — the underlying event stream is
  accessible through the API, not just the projected
  current state

The two audiences are equal by design, not by accident.

Relationship to CORTEX

CORTEX is the primary AI consumer of LOGOS data. The
quality of CORTEX's knowledge is directly bounded by the
fidelity of what LOGOS exposes through its API. A
first-class API is not a nice-to-have for CORTEX — it is
a hard dependency.

Acceptance

A post is created with metadata and a state transition.
A human views the full post including metadata and history
in the UI. An API call to the same post returns identical
content, metadata, history, and state — nothing omitted.
The event log for that post is accessible via API with
full fidelity.

Sources

• LOGOS → Requirements → Arbitrary post metadata (metadata must be API-accessible at full fidelity)
• LOGOS → Requirements → Event-sourced storage as the foundation (event log accessibility is a consequence of this foundation)