AcreFrame
Menu
Northline Biological Production DemoElevated Review Mode2026-05-10 08:42 ETReadiness 72/100Open Exceptions: 18QA Holds: 4Paused Batch Moves: 3Human Reviews Pending: 9Facility Watch: Elevated

System Architecture

Architecture for the operating system under regulated biological production.

AcreFrame is being built as a human-reviewed backend layer that connects facility records, rooms, batches, lots, QA holds, packet evidence, facility signals, owners, timestamps, distribution commitments, cost pressure, and decision memory.

Decision-support only. Human review required. No compliance certification, legal/regulatory advice, cultivation guidance, or autonomous facility control. Demo data is fictional.

Evidence Pipeline

From facility evidence to diligence readiness

AcreFrame treats evidence as a connected graph, not a folder of files. Each stage produces output that becomes input for the next. The pipeline below is illustrative — it shows how raw records become reviewable, linkable, and auditable.

Facility Evidence Pipeline

Fictional pipeline stages. Illustrative input/output only.

RecordsImplemented

In: Batch logs, SOP timestamps

Out: Structured record stream

NormalizationImplemented

In: Raw facility exports

Out: Canonical event schema

Entity GraphImplemented

In: Batch ID + lot + room

Out: Linked object graph

Failure Chain EngineImplemented

In: Drift signal + QA note

Out: Suspected root chain

Review QueueImplemented

In: Flagged evidence packet

Out: Owner-assigned review

Packet ReadinessImplemented

In: Section completeness map

Out: Readiness score + gaps

Decision LedgerImplemented

In: Attested human decision

Out: Immutable decision log

Diligence OutputImplemented

In: Reviewed evidence graph

Out: Diligence artifact set

Human-reviewed. Decision-support only.

Core v0 Scaffold

Core v0 is implemented as a typed demo scaffold

AcreFrame Core v0 ships with 18 typed entities, deterministic fictional demo data, a read-only repository layer, and five public API demo routes. The architecture below shows the full target stack; Core v0 implements the entity graph, signal event ledger, review queue, packet readiness, and decision memory layers as demo scaffold. Inspect the live scaffold →

7/10

Layers with scaffold objects

5

Public API demo routes

3

Layers still roadmap-only

0

Live customer integrations

Stack

Ten layers from evidence to diligence

Dashboards are presentation layers. AcreFrame is being built as an operating memory layer. The value is not the screen. The value is the connected evidence graph underneath the screen.

01

Record Intake

Implemented

Ingest exports, screenshots, spreadsheets, logs, and document samples from facility operations.

Inputs

Batch logs, QA notes, packet samples, inventory exports, facility reading screenshots, labor handoff notes, SOP excerpts, distribution timing records.

Outputs

Raw evidence files with source attribution and ingestion timestamp.

Humans decide what to send. AcreFrame does not access systems without explicit export.

02

Evidence Normalization

Design Partner

Parse heterogeneous inputs into a consistent schema with standardized timestamps, units, and completeness flags.

Inputs

Raw evidence files from Record Intake.

Outputs

Normalized evidence records with parsed timestamps, unit conversions, and gap flags.

Normalization rules are tuned per design partner. No automatic judgment of record validity.

03

Entity Graph

Implemented

Link facilities, rooms, zones, batches, lots, QA holds, packet sections, owners, and signals into a traversable graph.

Inputs

Normalized evidence records.

Outputs

Linked entity graph with relationship confidence scores and unresolved linkage gaps.

Entity links are suggestions. Humans confirm or correct relationships during review.

04

Failure Chain Engine

Demo

Detect connected failure chains across batch movement, QA holds, packet gaps, facility signals, labor queues, inventory aging, and cost pressure.

Inputs

Entity graph with temporal and spatial context.

Outputs

Failure chain paths with severity, dependency depth, and affected surfaces.

Failure chains surface connections for review. They do not predict the future or certify risk.

05

Review Queue

Design Partner

Score and route review items to qualified operators based on severity, age, dependency count, and owner clarity.

Inputs

Failure chains, entity gaps, and aging signals.

Outputs

Prioritized review queue with owner assignment, evidence links, and review windows.

Queue prioritization is decision-support only. Humans decide what to review and when.

06

Packet Readiness Engine

Design Partner

Map packet section completeness to batch movement gates, COA chains, and export readiness.

Inputs

Packet sections, COA records, batch state, and QA hold status.

Outputs

Packet gap matrix, readiness score, and blocked-gate attribution.

Packet readiness is a decision-support artifact. It does not certify compliance or guarantee inspection outcomes.

07

Decision Ledger

Design Partner

Preserve who reviewed what, what evidence was attached, what remained unresolved, and what decision was recorded.

Inputs

Review outcomes, human decisions, and evidence references.

Outputs

Immutable decision log with reconstructable audit trail per batch and per facility cycle.

Decisions are recorded by humans. AcreFrame preserves memory; it does not make decisions.

08

Terminal Surfaces

Implemented

Render operating memory as command-center surfaces for operators, supervisors, and diligence reviewers.

Inputs

Entity graph, failure chains, review queue, packet readiness, decision ledger.

Outputs

Terminal demo surfaces: Batch Movement Desk, QA Hold Queue, Packet Readiness Desk, Facility Signal Watch, Labor Load Board, Inventory/Distribution Desk, Cost Pressure Desk, Decision Log.

Terminal is a read-only decision-support surface. No control commands are issued.

09

Diligence Outputs

Implemented

Generate exportable readiness reviews, exception taxonomies, and operating-loop maps for investor and operator diligence.

Inputs

Decision ledger, packet readiness, failure chains, and entity graph snapshots.

Outputs

Facility Loop Review, Sample Report, operating risk taxonomy, batch dependency graph.

Diligence outputs are human-reviewed artifacts. They do not replace licensed professional review.

10

Read-Only Signal Bridge

Roadmap

Ingest facility signals from environmental sensors, equipment logs, and energy intervals for contextual enrichment.

Inputs

Temperature, humidity, VPD, pressure, equipment runtime, environmental alarms, energy intervals, water-system logs where available.

Outputs

Normalized signal events linked to rooms, batches, and review queues.

Signal Bridge is read-only by design. It does not control HVAC, irrigation, lighting, dosing, QA release, movement, or shipment.

Capabilities

Current vs Roadmap Matrix

CapabilityStatusData RequiredHuman RoleBoundary
Facility Loop ReviewImplementedOperator records, exports, screenshots, walkthrough notes.Operator provides evidence. AcreFrame maps failure chains and produces readiness review.Does not certify compliance or replace licensed review.
Terminal DemoImplementedFictional demo dataset (Northline Cultivation).Visitor explores fictional surfaces. No real customer data.Fictional demo data. Not customer data.
Batch Object ModelImplementedBatch logs, lot records, movement records, owner assignments.Operator confirms batch-linkage suggestions.Does not approve batch movement or release QA holds.
QA Hold QueueImplementedQA hold records, aging history, owner notes, retest logs.Supervisor reviews queue priority and assigns owners.Does not release holds or certify QA outcomes.
Packet Readiness MatrixImplementedPacket sections, COA files, attestation records, export requirements.Compliance lead reviews gaps and decides what to complete.Does not certify compliance or guarantee inspection outcomes.
Decision LedgerImplementedReview outcomes, human decisions, evidence references.Operators record decisions. AcreFrame preserves memory.AcreFrame preserves what humans decided. It does not decide.
Read-Only IntegrationsRoadmapLawful exports from seed-to-sale, task tools, inventory systems, environmental logs.Operator exports data. AcreFrame ingests and normalizes.Does not replace regulated systems or access live databases without consent.
Signal BridgeRoadmapEnvironmental sensor streams, equipment event logs, energy intervals.Facilities engineer reviews signal context linked to batches and holds.Read-only signal ingestion. No equipment control.
Multi-Facility MemoryNot OfferedCross-facility entity graphs, shared taxonomy, consolidated decision logs.Corporate operations lead reviews multi-facility operating memory.Not currently offered. Future infrastructure stage only.
Diligence OutputsImplementedFacility Loop Review artifacts, decision logs, packet matrices.Investor or operator reviews human-prepared diligence artifacts.Does not replace licensed legal, financial, or regulatory diligence.

Thesis

Why backend architecture matters

Dashboards are presentation layers. AcreFrame is being built as an operating memory layer. The value is not the screen. The value is the connected evidence graph underneath the screen. When a QA hold, a humidity excursion, and a missing packet section all point to the same batch, the connection matters more than any individual alert. That connection requires durable entities, normalized evidence, and a traversable graph — not just prettier charts.

Human Authority Boundary

Decision-support only

What AcreFrame does and does not do

AcreFrame may

  • Ingest records and normalize evidence
  • Link facilities, rooms, batches, lots, holds, packets, and signals
  • Surface failure chains and score review pressure
  • Route human review to qualified operators
  • Preserve decision memory across shifts and cycles
  • Prepare packet-readiness artifacts
  • Show read-only facility signal context where available
  • Generate fictional demo packet previews

AcreFrame does not

  • Certify compliance or guarantee inspection outcomes
  • Control equipment or environmental setpoints
  • Prescribe cultivation, pesticide, nutrient, or watering actions
  • Provide formulation, extraction, or dosing instructions
  • Release QA holds or approve batch movement
  • Approve shipments or authorize distribution
  • Replace licensed professionals or legal counsel
  • Promise specific yield, potency, savings, or margin outcomes

Operating Loop

How records become memory

The operating loop is not a linear workflow. It is a cycle: records normalize into entities, entities link to decisions, decisions produce memory, and memory informs the next record. AcreFrame preserves state across every rotation.

Records
Normalization
Entity Linking
QA Review
Packet Check
Batch Movement
Distribution
Decision Memory

Fictional operating loop. Decision-support only. Human review required at every stage.

Object Model

What a connected batch object looks like

In AcreFrame, a batch is not a row in a spreadsheet. It is an object with linked signals, QA state, packet sections, owners, decisions, and cost exposure. The view below is read-only and uses fictional demo data.

Object Anatomy · Fictional Demo Data

read-only
batch
AF-1042
room
Zone 4 Flower
lot
LOT-GUM-21
stage
Transfer
QA state
HLD-092 · 52h
linked signals
RH drift above baseline
packet sections
Testing (blocked), Harvest (partial)Blocked
owners
N. Lopez (QA), S. Rivera (Compliance)
decisions
DEC-7708 (unresolved)
distribution dependency
MX-440
cost exposure
Packaging rework + HVAC delta — elevatedElevated
unresolved blockers
3

Boundary

Read-only object view. Human review required.

Evidence Stack

What evidence looks like at the facility level

Evidence completeness is not binary. AcreFrame tracks partial, complete, and missing evidence per batch and per packet section. The stack below shows illustrative status for a fictional facility.

Facility Evidence Stack

Fictional demo evidence. All labels are illustrative.

Batch logs

Seed-to-sale export

complete

QA notes

Observation + review record

partial

Packet samples

Section attestation

missing

Inventory exports

Lot-level aging table

complete

Facility readings

Environmental signal log

partial

Distribution timing

Manifest + commitment

complete

Evidence stack shows illustrative status only. Human review required for all compliance decisions.

Design partner

Map one broken operating loop.

If you are a licensed operator with real records and operational complexity that has outgrown spreadsheets, request a Facility Loop Review.

AcreFrame State · Honest product boundary

Software only follows where recurring pain is validated.

Implemented

Facility Loop Review
Terminal Demo
Batch Object Model
QA Hold Queue
Packet Readiness Matrix
Decision Ledger
Diligence Outputs

Design Partner

Evidence normalization
Review routing
Packet-readiness engine tuning

Roadmap

Read-only integrations
Read-only Signal Bridge
Multi-facility memory

Not Offered

Autonomous control
Compliance certification
Legal/regulatory advice
Cultivation treatment recommendations
Pesticide/nutrient/watering prescriptions

Decision-support only. No autonomous control, compliance certification, legal/regulatory advice, cultivation guidance, pesticide guidance, treatment instructions, nutrient or irrigation prescriptions.

AcreFrame does not provide legal advice or regulatory advice. AcreFrame does not provide pesticide guidance, cultivation treatment instructions, irrigation or watering prescriptions, or nutrient dosing prescriptions. No pesticide. No cultivation treatment prescriptions. AcreFrame does not autonomously control HVAC, irrigation, lighting, water systems, dosing, QA release, remediation, harvest, packaging, shipment, or compliance decisions. No autonomous control. Qualified human review required. AcreFrame does not provide compliance certification. No compliance certification. AcreFrame does not guarantee compliance, yield, potency, margin, savings, audit outcomes, or risk elimination. Licensed operators only where lawful.