Fly larvae composting is an emerging waste treatment alternative with great potential to increase revenue from food waste management. For wider implementation, fly larvae composting has to be evaluated in comparison with conventional systems, based on direct greenhouse gas (GHG) emission data for the treatment process, which are currently limited. This study evaluated direct emissions of CO2, CH4, N2O and NH3 from composting of food waste using black soldier fly (BSF) larvae (Hermetia illucens). Use of BSF larvae-associated bacteria in 7-day pre-treatment and seeding at larvae treatment start were evaluated and compared to larvae treatment without bacteria addition. The treatments were performed in a set of 14-day laboratory-scale experiments. Mean substrate reduction was 49 +/- 8% and bioconversion ratio was 24 +/- 8% (both dry matter basis). Direct GHG emissions from the fly larvae treatment process were generally very small, with emissions of CH4 and N2O equivalent to 0.38 kg CO2-equivalents per ton food waste treated assuming global warming potential over 100 years, while mean total CO2 emissions were 96 g CO2 per kg food waste treated. Additional emissions could be expected to occur in the pre-treatment process, which did not provide any significant improvement in bioconversion ratio or reduction in total GHG emissions during treatment. Similarly, use of BSF larvae-associated bacteria did not significantly improve process efficiency. No NH3 emissions were detected, as reflected in total N mass balance over the treatment cycle. The results show that total direct GHG emissions from food waste treatment by fly larvae composting are lower than those from conventional food waste treatment, and that pre-treatment and seeding of food waste with BSF larvae-associated bacteria do not further reduce total GHG emissions. (C) 2019 Elsevier Ltd. All rights reserved.