02 · MyCelAI
Built for agriculture, climate, and energy at the edge: fields, sites, and grids that cannot wait on a cloud round-trip. Exportable reporting when the connection isn’t there.
At the edge
Most AI assumes the network is there. Operators in the field, on a site, or on a grid often cannot. MyCelAI is local-first and offline-first by default: inference and reporting run where the work is, then export when a link returns.
The people who carry the decision (agronomists, climate operators, energy managers) stay in control. The system augments them. It does not replace accountable judgment.
Where it runs
One conductor. Three edge environments that cannot wait on the cloud.
Fields and operations
Local advisories on moisture, nutrients, and operations that still produce a report when the tower is down. Built for people who make decisions in the field, not in a dashboard that needs five bars.
Offline-first · Exportable · Human-in-the-loop
Sites and monitoring
Edge processing for environmental and carbon-related reporting. Data stays local until it is ready to export. Traceable outputs, not a black-box score.
Local · Verifiable · Exportable
Grids and facilities
Monitoring and decision support at facilities and distribution points that cannot wait on a round-trip. The operator stays in charge of what gets acted on.
Edge · Offline-capable · Operator in charge
Standard
Work happens on-device and on-site. Cloud is an export path, not a requirement for the system to function.
Reports are structured so an operator, auditor, or buyer can see what was measured, what was inferred, and what is labeled as an estimate.
MyCelAI organizes signal for experts. It does not take the decision away from them.
Science, below the product
MyCelAI is software. In some deployments it listens to electrical activity in living mycelium networks and turns those waveforms into structured signal for operators at the edge. That research layer is not the product headline. It is how some of the sensing gets done.
LaRocco et al. (2025), PLOS ONE. “Signal interpretation in fungal mycelial networks” shows that living mycelium networks exhibit measurable, patterned electrical activity in response to environmental stimuli. That paper is a foundation for signal interpretation, not a claim that the product is a mushroom. LaRocco et al. (2025). PLOS ONE. Foundation paper for mycelial signal interpretation used in some MyCelAI deployments.
Where biological sensing is used, the conductor treats electrical activity as a signal to classify and report (moisture, temperature, chemical gradients, mechanical disturbance) with the same audit posture as the rest of the stack.
Signal processing is designed to run locally. A site that loses connectivity still collects, and still produces an exportable report when the link returns.
Not every Conductor deployment is a biological sensor network. The product is the local-first intelligence layer. The substrate varies with the site.
If you run agriculture, climate, or energy operations that cannot wait on a cloud round-trip, talk to us.
We work with a few teams at a time.
VedAGI. Intelligence for Humanity.