Hojjat, AliAliHojjatHaberer, JanekJanekHabererLandsiedel, OlafOlafLandsiedel2026-07-312026-07-312026-07-30International Journal of Computer Vision 134 (8): 374 (2026)https://hdl.handle.net/11420/64165The rapid growth of camera-based IoT devices demands the need for efficient video compression, particularly for edge applications where devices face hardware constraints, often with only 1 or 2 MB of RAM and unstable internet connections. Traditional and deep video compression methods are designed for high-end hardware, exceeding the capabilities of these constrained devices. Consequently, video compression in these scenarios is often limited to M-JPEG due to its high hardware efficiency and low5 complexity. This paper introduces MCUCoder , an open-source adaptive bitrate video compression model tailored for resource-limited IoT settings. MCUCoder features an ultra-lightweight encoder with only 10.5K parameters and a minimal 350KB memory footprint, making it well-suited for edge devices and MCUs. While MCUCoder uses a similar amount of energy as M-JPEG, it reduces bitrate by 55.65% on the MCL-JCV dataset and 55.59% on the UVG dataset, measured in MS-SSIM. MCUCoder supports adaptive-bitrate transmission by generating a latent representation sorted by importance, allowing the transmitted data to be adjusted according to available bandwidth. We further show that MCUCoder compression outperforms M-JPEG compression when evaluated on downstream AI tasks, including image classification, object detection, and image captioning. Source code available at https://github.com/ds-kiel/MCUCoder .en1573-140520268Springerhttps://creativecommons.org/licenses/by/4.0/Deep video compressionInternet of thingsAdaptive bitrate encodingComputer Science, Information and General Works::006: Special computer methodsJournal Article2026-07-3110.1007/s11263-026-02953-610.15480/882.17731