Root of the integrated Master Grid platform. This service is the first contact surface for every event, glyph expression, document, or signal entering the system. It implements the Master Grid 7-layer execution contract at the boundary, classifies inputs against the NLCA 10-field metadata schema, parses glyph DSL expressions natively via the Glyphic Codex parser, and dispatches to downstream services through a central event bus.
The architecture is derived from a single visual: the Input Processing Layer keynote-render. Roughly fifty translucent prismatic columns of varying heights stand in a hexagonal cluster on a fractured crystalline base; a warm amber glow emanates from the center, visible through the translucent forms.
Visual element → code element:
| Visual | Code |
|---|---|
| ~50 translucent columns | N parallel ingest channels (router.py) |
| Variable column heights | Variable processing depth per channel (per-channel inspector pipelines) |
| Central warm-amber glow | Central dispatcher / event bus (dispatcher.py) |
| Hexagonal lattice base | Persistent substrate — SQLite audit + state |
| Translucent material | Observability promise — every step is introspectable; no hidden state |
┌──────────────────────────────────────────┐
inbound ──▶ │ SEED CRYSTAL — INPUT PROCESSING LAYER │ ──▶ outbound
│ (this service — root of the platform) │ to NLCA, Glyphic Codex,
└──────────────────────────────────────────┘ AI_Memory_Core, etc.
│
▼
Master Grid 7-layer execution contract
L1 Baseline → L7 Response (every input crosses this)
Every input that crosses this service is annotated with its Master Grid position. Layers L1–L6 happen before dispatch; L7 happens at response.
| Layer | What this service does |
|---|---|
| L1 Baseline | Tag the input with a doc_id matching ^[A-Z]{3,4}-[A-Z]{3,6}-[0-9]{3}$; record lifecycle stage |
| L2 Coordination | Define done-state (RECEIVED → CLASSIFIED → DISPATCHED → ACKNOWLEDGED) |
| L3 Fact Work | Inspect the input — channel-specific inspectors run concurrently |
| L4 Design Work | Route to handler via router.py; choose smallest defensible dispatch |
| L5 Integration | Hand off to the central dispatcher for downstream wiring |
| L6 Governance | Validate schema, scrub PII candidates, enforce rate limits, immutable audit log |
| L7 Response | Return outcome envelope: {outcome, files, rationale, validation, risks, next} |
A first-class endpoint accepts glyph DSL expressions directly:
curl -X POST http://localhost:8000/input/glyph \
-H "Content-Type: application/json" \
-d '{"expression": "📥 → ⚙️ → 📤"}'
# → {"ast_function_form": "input() → process() → output()",
# "lifecycle": "DISPATCHED",
# "doc_id": "INP-GLYPH-001",
# "master_grid_layer": "L4"}The parser is the same glyph_dsl.py published in the Glyphic Codex DSL
repository (lineage from AI_Memory_Core), embedded here as
seed_crystal/parsers/glyph.py.
# Install deps
pip install -r requirements.txt
# Run the API
uvicorn seed_crystal.app:app --reload
# Swagger UI
open http://localhost:8000/docs
# Run tests
pytest -v| Method | Path | Purpose |
|---|---|---|
| POST | /input/event |
Generic input event — accepts any payload + channel tag |
| POST | /input/glyph |
Glyph DSL expression intake (parses + dispatches) |
| GET | /input/lifecycle/{doc_id} |
Inspect lifecycle state of a prior input |
| GET | /input/audit |
Read the immutable audit log (paginated) |
| GET | /healthz |
Liveness check |
| GET | /docs |
Swagger UI |
| GET | /openapi.json |
OpenAPI schema |
Seed Crystal Input Processing Layer/
├── README.md ← this file
├── ARCHITECTURE.md ← Master Grid mapping detail
├── LICENSE ← MIT
├── .gitignore
├── requirements.txt
├── pyproject.toml
├── seed_crystal/
│ ├── __init__.py
│ ├── app.py ← FastAPI app (entry point)
│ ├── master_grid.py ← 7-layer contract definitions
│ ├── models.py ← Pydantic event models + NLCA 10-field schema
│ ├── lifecycle.py ← FSM: RECEIVED → CLASSIFIED → DISPATCHED → ACK
│ ├── router.py ← Channel → handler routing
│ ├── dispatcher.py ← Central event dispatcher (the golden glow)
│ ├── audit.py ← Immutable audit log
│ ├── parsers/
│ │ ├── __init__.py
│ │ ├── glyph.py ← Glyph DSL parser (from Glyphic Codex)
│ │ └── glyph_codex.py ← GLYPH_TO_FUNCTION map (stub mirror of AMC)
│ └── inspectors/
│ ├── __init__.py
│ └── default.py ← Default inspector (extensible per-channel)
├── tests/
│ ├── __init__.py
│ ├── test_master_grid.py
│ ├── test_lifecycle.py
│ ├── test_dispatcher.py
│ ├── test_glyph_parser.py
│ └── test_app.py ← End-to-end FastAPI TestClient tests
└── docs/
└── SEED_GENESIS.md ← captures the seed-crystal image origin
| Companion project | How Seed Crystal integrates |
|---|---|
| Glyphic Codex DSL (canon #16) | Embeds glyph_dsl.py as a first-class parser at /input/glyph |
| AI_Memory_Core (canon #7) | Audit-log entries can be exported to AMC's document_archive/ format |
| Neural Lattice (canon #10) | Output envelope conforms to NLCA 10-field metadata schema; can POST directly to NLCA /api/documents |
| 7 Layer Architecture Template Pack (canon #11) | Master Grid 7-layer contract is the runtime instantiation of the Phase Pack |
| iOS Compliance Framework (canon #14) | L6 Governance hook can call iOS-compliance validators when inputs target mobile |
| SDLC Framework (canon #13) | Audit log is structured for SDLC-style traceability matrix consumption |
| Hub and Spoke AI Chatbot (canon #15) | Seed Crystal IS the hub; downstream services are the spokes |
MIT — see LICENSE.
Shannon Brian Kelley — AI Orchestrator Architect. Built in collaboration with Claude AI (Anthropic) under file-system-bound persona protocol; co-creator role attributed as "Archivist of Wisdom." Genesis image declared seed crystal 2026-05-02.