System architecture & data flow

A monitoring layer built
for every grow room

MycoFarm connects directly to your sensors and controls — not on top of them. From raw IoT signals to automated harvest decisions, here's exactly how it works.

Request a demo See capabilities
scroll
<60s
Alert response time from sensor read
80+
Supported IoT protocols & sensor models
100%
On-premise — no grow data leaves your facility
5ms
Sensor polling interval for critical paths

A reasoning layer that sits
across your entire grow operation

MycoFarm connects directly to your sensors and environmental controls — not on top of them. Our engine reads from your IoT devices, HVAC controllers, and substrate logs without any data replication lag.

01

Sensor integration

Connect temperature, humidity, CO₂, light, and weight sensors via native integrations with major IoT protocols — no custom hardware required.

02

AI-driven analysis

Purpose-built models that understand mushroom biology, flush cycles, contamination vectors, and species-specific thresholds — not just generic sensor data.

03

Automated action & alerts

Every anomaly triggers a configurable response: HVAC adjustment, grower notification, or harvest flag — with a full log of what changed and why.

04

Continuous learning

Models adapt to your facility over time — refining baselines, seasonal patterns, and species tolerances as each flush cycle adds to the knowledge base.

MYCO
FARM
ENGINE

From raw signal to
actionable decision

Every data point travels through four processing stages before it reaches a grower or triggers an automated response.

Sensor Layer

Temp, RH, CO₂, light, weight, camera

ingest

Normalization

Protocol parsing, calibration, outlier filtering

enrich

AI Inference

Anomaly detection, yield forecast, species models

act

Response & Log

HVAC trigger, alert dispatch, audit trail

Each stage in detail

Select a phase to see how MycoFarm handles every step of the cultivation intelligence loop.

PHASE 01 Sensor Integration & Ingestion

MycoFarm supports over 80 sensor models and IoT protocols out of the box. Devices are discovered automatically on your local network, calibrated against known baselines, and begin streaming within minutes of installation. No custom firmware or hardware bridges required.

Protocols

MQTT, Modbus RTU/TCP, BACnet, Z-Wave, Zigbee, OPC-UA, and direct REST/webhook APIs from major sensor manufacturers.

Sensor types

Temperature, relative humidity, CO₂, VPD, PAR light, substrate weight, airflow velocity, and IP camera streams for computer vision.

Polling rates

Standard paths poll every 60 seconds. Critical paths — contamination cameras and temperature runaway — operate at 5ms intervals with dedicated processing threads.

PHASE 02 Normalization & Contextualization

Raw sensor readings arrive in dozens of formats and units. The normalization layer parses each protocol, applies device-specific calibration offsets, filters electrical noise and dropout artifacts, and tags every reading with room ID, species profile, flush cycle number, and substrate batch — giving the AI layer full environmental context for every data point.

Calibration

Automated drift correction runs weekly. Growers can trigger manual recalibration events that propagate corrections backward through the historical log.

Tagging

Each reading is enriched with grow room, species, substrate recipe, inoculation date, and flush number — creating a fully contextualized time-series record.

Storage

All normalized data is stored on-premise in a compressed time-series database. Retention is configurable — typically two years at full resolution.

PHASE 03 AI-Driven Analysis & Inference

Purpose-built models trained on thousands of grow cycles run continuously against the normalized stream. Unlike generic anomaly detectors, MycoFarm's models understand mushroom biology: pin initiation signals, mycelium run phases, contamination precursors, and the environmental fingerprint of each species. Inference runs entirely on your local hardware — no cloud API calls required.

Contamination detection

Environmental signatures and computer vision classify bacterial and fungal contamination 48–72 hours before visual symptoms — early enough to isolate affected blocks before spread.

Yield forecasting

Flush weight predictions combine sensor history, substrate composition, inoculation timing, and historical flush performance for each species and room combination.

Adaptive baselines

Species-level thresholds refine automatically as your operation accumulates data. New rooms reach steady-state model accuracy within three to five flush cycles.

PHASE 04 Automated Response & Human Escalation

Every inference output triggers a configurable downstream action. Low-severity events update the dashboard and append to the audit log. Medium-severity events dispatch SMS, email, or app push notifications to the responsible grower. High-severity events — contamination risk, temperature runaway, or harvest window expiry — can directly trigger HVAC adjustments, irrigation changes, or equipment shutoffs via your existing control bus.

Notification channels

SMS, email, Slack, app push, and webhook endpoints. Escalation chains ensure critical alerts reach on-call staff even outside business hours.

HVAC control

Direct integration with BACnet and Modbus HVAC controllers lets MycoFarm adjust temperature, humidity, and CO₂ targets automatically within grower-defined safe ranges.

Audit trail

Every automated action is logged with timestamp, triggering inference, confidence score, and grower acknowledgement — HACCP and organic certification ready.

PHASE 05 Continuous Learning & Model Refinement

MycoFarm improves with every flush cycle. Outcome data — actual yields, contamination incidents, harvest quality ratings — feeds back into the model training loop automatically. Growers can label events directly in the dashboard, and those labels prioritize retraining on the cases that matter most to your operation. No manual model management required.

Feedback loop

Harvest weight actuals, contamination confirmations, and grower-labeled alerts retrain species models on a rolling weekly cycle without service interruption.

Model versioning

All model versions are stored with performance metrics. Rollback to any prior version in one click if a retrain produces unexpected behaviour.

Baseline drift

Seasonal environmental shifts, substrate recipe changes, and new strain introductions are detected and incorporated automatically without grower intervention.

Connects to the hardware
and software you already use

Environmental Sensors

Inkbird, SensorPush, Govee, Aranet, and 60+ additional sensor brands with native drivers.

MQTT · Zigbee · BLE

HVAC & Climate Control

Inkbird IHC-200, Johnson Controls, Honeywell, and any BACnet or Modbus-compatible controller.

BACnet · Modbus RTU/TCP

Computer Vision

IP cameras via RTSP/ONVIF. Plug in any PoE camera to enable contamination and pin-set detection.

ONVIF · RTSP

Substrate & ERP Systems

CSV import, REST API, and direct database connectors for custom farm management and ERP software.

REST API · Webhooks

Alert Channels

SMS, email, Slack, PagerDuty, and custom webhooks. Configure per-room, per-species escalation chains.

Slack · SMS · Webhooks

Compliance & Audit

Automatic HACCP log export in PDF and CSV. Pre-formatted templates for USDA Organic, Canada Organic, and EU Organic audits.

HACCP · Organic Cert

Dashboard & Reporting

Embedded web dashboard accessible on any device. REST API for building custom reports or integrating with existing business intelligence tools.

Web UI · REST API

On-Premise Storage

Time-series and relational data stored locally on your hardware. No cloud dependency — runs fully air-gapped if required.

Local · Air-gapped

Live in days,
not months

Day 1

Hardware discovery & sensor onboarding

MycoFarm scans your local network, discovers compatible devices, and walks you through pairing any remaining sensors. Most farms complete sensor setup in under four hours with no external help required.

Day 1–2

Room profiles & species configuration

Configure each grow room with its dimensions, ventilation type, and active species. Pre-loaded cultivation profiles for oyster, shiitake, lion's mane, reishi, king trumpet, and nine other species populate baseline thresholds automatically.

Day 2–3

HVAC & control integration

Connect MycoFarm to your HVAC controllers via BACnet or Modbus. Define safe automation ranges — the system will only act within grower-approved bounds, and every automated adjustment is logged and reversible.

Day 3–5

Baseline calibration & alert tuning

During the first 72 hours, MycoFarm establishes environmental baselines for each room. Alert thresholds are auto-suggested based on species profiles and can be refined by growers before going live.

Day 5+

Full operational monitoring

MycoFarm is live. From this point, every flush cycle adds data to the model training loop, improving forecast accuracy and contamination detection precision with each successive harvest.

Ready to grow smarter
with precision cultivation?

Join farms using MycoFarm to increase yields, reduce contamination losses, and run cleaner operations at any scale.