How Do Electric Composters Work, and Is the Output Compost?

How much of the effect is Kobold [starter material] itself versus the environment it creates?

The two are closely connected.

Kobold provides the biological starting capacity. It introduces a prepared microbial community that can begin working across a broad range of normal kitchen waste, rather than leaving the system entirely dependent on whichever microorganisms happen to arrive with the food.

It does not sterilise the input or permanently exclude every microorganism already present. Food brings its own microbial population into the chamber, and the community inside Terra II continues to change with the feeding pattern.

Terra II provides the conditions that make this biological activity more consistent: oxygen, moisture management, temperature support and regular contact between the microbes and the food.

A capable microbial culture cannot perform well for long if the material becomes waterlogged, compacted, oxygen-limited or excessively dry. Equally, a well-controlled environment without a reliable starting community would usually produce a less predictable start-up and a greater variation between different food types.

The simplest way we describe it is: Kobold provides the biology; Terra II provides the conditions that allow it to work consistently.

Where does Terra II sit in relation to mesophilic, thermophilic and curing stages?

A traditional batch compost pile may move broadly from a mesophilic phase into a thermophilic phase, then cool and cure.

Terra II does not move through those stages as one single mass. Because it is continuously fed, different parts of the material can be at different stages at the same time.

Fresh, room-temperature food may pass through a local mesophilic transition after it is added. The established active matrix operates mainly in a higher-temperature biological range, generally around 45–55°C. Older material is meanwhile moving into a lower-rate stabilisation phase.

We would not claim that every part of the chamber remains at exactly the same temperature, or that every particle reaches the same degree of maturity at the same moment. Local moisture, porosity, food composition and airflow all affect the process.

The harvested material is a biologically processed, concentrated compost output. It is not simply dried or ground food waste. At the same time, it should not be treated as uniform, inert potting soil.

For normal garden use, we recommend mixing approximately one part GEME output with eight parts soil and covering it with at least 5 cm of soil. We do not recommend using fresh output as the sole growing medium, particularly for seedlings or salt-sensitive plants.

A continuous system does not have one single moment when every particle becomes equally mature.

Have you tested CO₂ respiration or volatile solids reduction?

We agree that respiration, oxygen-uptake and volatile-solids testing would provide stronger evidence of biological stabilisation than appearance, heat or weight reduction alone.

The data we have published so far has focused mainly on output composition, including organic matter, nutrients and pH, together with instrumented observations of the decomposition process.

We have not yet published a standardised CO₂ respiration, oxygen-uptake or volatile-solids-reduction dataset for Terra II output that we would ask an independent reviewer to treat as conclusive evidence of stability.

We do have clear process evidence of active biological decomposition, but we do not treat process heat, visible breakdown or mass loss as substitutes for a standardised endpoint test. Weight loss alone cannot distinguish water evaporation from biological conversion of organic matter.

Continuous systems also create a sampling challenge. A sample taken shortly after a large food addition will not have the same material age as a sample taken from the established compost matrix. A meaningful study therefore needs a defined feed profile, sampling position, material age and conditioning period.

Standardised respiration or oxygen-uptake testing, together with volatile-solids analysis under a controlled feeding and sampling protocol, is one of the areas we are prioritising for the next validation stage.

Has salinity or EC come up in testing?

Yes. It is an important issue in any system that concentrates food-derived material into a relatively small output volume.

Salts do not biodegrade in the way carbohydrates, proteins and fats do. As organic carbon is metabolised and water leaves the system, soluble minerals can become more concentrated. EC may therefore rise even when the biological process itself is operating normally.

Terra II is designed to handle the normal variation found in household leftovers, including reasonable amounts of salty or oily food. That should not be interpreted as unlimited tolerance for brine, cooking oil, concentrated sauces or repeated high-salt inputs.

Our practical advice is to: drain excessive brine, soup or sauce before adding the solid food; avoid pouring cooking oil directly into the chamber; divide unusually salty or fatty food across several feedings; maintain a varied input mix rather than repeatedly feeding one concentrated waste type; and mix the harvested output with soil at the recommended ratio rather than using it undiluted.

Salt and fat also create different problems. Salt is mainly a concentration and plant-use issue. Excess fat can coat particles, restrict air and water transfer, and temporarily slow aerobic decomposition.

We are now looking at EC across defined feeding profiles rather than treating one isolated output sample as representative of every household diet.

Salt does not biodegrade, so the responsible answer is good input management and appropriate soil dilution.

What did R&D find mattered more or less than expected: the culture or environmental control?

One of the clearest lessons from our R&D work was that strain selection alone could not deliver consistent performance.

The microbial culture matters, but the physical microenvironment is more complex than it first appears. Two systems can show a similar average temperature while having very different moisture distribution, airflow and biological performance.

Food releases water as it decomposes. Material can settle, become denser or create local areas where air no longer moves effectively. This means that moisture distribution, porosity, mixing resistance and the actual route taken by the air can matter as much as the headline temperature.

We also learned that holding one exact temperature is less important than maintaining a suitable operating window and recognising when the system is moving away from it.

More heat is not automatically better. Heat can support biological activity, but it cannot replace oxygen, moisture balance or effective contact between the microbes and the food.

Particle size follows the same principle. Reducing size can improve microbial access, but excessive breakdown of wet food into a dense slurry can reduce effective pore space and restrict oxygen transfer. Terra II therefore uses controlled mixing rather than treating fine grinding as the objective.

The culture provides metabolic capability. Environmental control determines how reliably that capability is expressed.

What is GEMEBrain reading to decide that a cycle is done?

Strictly speaking, GEMEBrain does not decide that the entire contents of Terra II are “done”, because Terra II is not a single-age batch.

It reads a combination of signals, including temperature behaviour, moisture-related conditions, changes in load or mass, mechanical resistance during mixing and the recent operating history of the machine.

The important point is that it looks at how those signals change together over time, rather than relying on one threshold or a fixed countdown.

Temperature can rise because of microbial activity, auxiliary heating or a change in room conditions. Mechanical resistance can change because the material is wetter, denser, more fibrous or unevenly distributed. Moisture-related signals also need to be interpreted in the context of what was recently added.

GEMEBrain uses these combined trends to adjust aeration, heat and mixing intensity. As the strong response associated with a recent addition begins to settle, the system can move from more active intervention towards a lower-intensity maintenance state.

That is an operational judgement about the condition of the composting bed. It is not a laboratory declaration that every particle has reached the same maturity.

GEMEBrain manages the process state; it does not replace laboratory maturity testing.

What is next on the testing roadmap?

Our next-stage work is focused on making the performance of the system more measurable, reproducible and independently verifiable.

The main priorities are: standardised CO₂ respiration or oxygen-uptake testing; volatile-solids reduction under defined feeding and sampling conditions; EC and salinity testing across representative household feed profiles; germination and plant-response testing at different compost-to-soil ratios; repeatability testing using defined mixed-food inputs rather than selected single ingredients; controlled comparisons between microbial inoculation and environmental control; and stress testing involving temporary high-moisture, high-fat and high-salt inputs.

We expect this work to combine internal instrumented trials with independent laboratory analysis. Internal testing is necessary for understanding how the control system responds in real time. Independent testing is better suited to verifying output properties against defined methods.

We also want to be more explicit about test boundaries. A result produced from one feed composition, one sampling time or one set of environmental conditions should not automatically be presented as a universal claim.

When we publish further results, our intention is to show the feeding assumptions, sampling method and limitations alongside the headline figures. It is less convenient than presenting one simple number, but it is the only way to build a credible evidence base for a household biological system.