# CoreAI

CoreAI is a filesystem-native foundation for building, publishing, and preserving operational context for AI-assisted work.

It is intentionally simple. CoreAI starts with readable files, explicit manifests, reproducible bundles, and forensic snapshots. It does not require microservices, Kubernetes, a vector database, autonomous agents, or a specific AI provider.

## Objectives

- Keep project context understandable by humans and usable by AI systems.
- Make runtime context auditable, versioned, and reproducible.
- Preserve operational continuity through snapshots, timelines, and evidence.
- Keep the architecture provider-independent and filesystem-native.
- Let domains grow progressively on top of a stable core.

## Philosophy

CoreAI follows five baseline principles:

- **Context First**: AI output is only useful when the input context is explicit and inspectable.
- **Filesystem Native**: the filesystem is the first source of truth.
- **Documentation Native**: Markdown and YAML are primary project artifacts.
- **Human Readability**: every important file must be readable without special tooling.
- **Provider Independence**: no model provider owns the architecture.

## General Architecture

CoreAI is organized around a small set of stable concepts:

- **Documents** describe intent, architecture, operations, and decisions.
- **Manifests** describe files, metadata, visibility, trust, and relationships.
- **Bundles** compose selected context for a specific use case.
- **Snapshots** preserve state at a point in time.
- **Exports** publish context in human-readable and AI-readable forms.
- **Domains** provide operational knowledge without changing the core.

## Repository Map

```text
docs/                 Official CoreAI v1 documentation
examples/             Minimal example manifests, bundles, snapshots, and exports
schemas/              YAML schemas for CoreAI data models
domains/HAS84/docs/   HAS84 origin story and operational proof notes
```

## Documentation Quick Start

Read these first:

1. [Foundations](docs/COREAI_FOUNDATIONS.md)
2. [Architecture](docs/COREAI_ARCHITECTURE.md)
3. [Filesystem Model](docs/COREAI_FILESYSTEM_MODEL.md)
4. [Manifest Model](docs/COREAI_MANIFEST_MODEL.md)
5. [Bundle Model](docs/COREAI_BUNDLE_MODEL.md)
6. [Forensic Model](docs/COREAI_FORENSIC_MODEL.md)
7. [Roadmap](docs/COREAI_ROADMAP.md)

## V1 Scope

CoreAI v1 is documentation-first. The goal is to stabilize the project language, filesystem layout, data models, and forensic assumptions before implementing runtime tooling.