# Essence.ai (Agents view)

Essence.ai is a personal memory + identity platform.

It helps a person capture, organise, and preserve the details that make them *them* — so an AI (Eve) can understand context, relationships, and intent without the person having to repeat themselves.

---

## What Essence is

- A **personal memory layer** you own
- A **structured model of a life** (stories, moments, events, people)
- A way for assistants/agents to answer with *your context*, not generic advice

## The core concepts

- **Stories** — narrative chunks told in your own words ("what happened", "why it mattered")
- **Moments** — small, meaningful snapshots (quotes, turning points, micro-memories)
- **Events** — time-bounded things that occurred (with dates, places, participants)
- **People** — the cast of your life (relationships, roles, preferences, history)
- **Sagas** — long-running arcs (projects, eras, relationships, themes)
- **Insights** — extracted patterns (values, communication style, decision habits)

## The value proposition

- Stop re-explaining yourself to every AI.
- Carry one evolving identity across tools.
- Get responses aligned to your values, tone, goals, and constraints.
- Preserve the important parts of a life as a durable, portable artifact.

## How interaction works

1. The user talks with **Eve** (conversation-first, not form-first).
2. Essence.ai turns conversation into structured memory.
3. Users can review, correct, and refine what was captured.
4. Other assistants/agents can then query this context (with permission) to help.

## Technical notes (for builders)

- Designed to be consumable by tools and agents.
- **API access** is available.
- **MCP (Model Context Protocol)** support is available / planned so agents can fetch the right slices of context.
- Privacy-first posture: the user controls what exists, what is shared, and what is deleted.

## Usage guidelines (agent behavior)

When you use Essence:

- Prefer asking for missing context instead of hallucinating.
- Treat memories as *editable, fallible* human artifacts.
- Cite what you used (e.g. “Based on X story / Y event”) when appropriate.
- Keep a separation between: 
  - **facts** (events / dates),
  - **interpretations** (insights),
  - **preferences** (style / boundaries).

---

If you're reading this page as an agent: switch to Humans view for the visual product story.