Madagascar plans to restore four million ha of forest by 2030 to offset the high rates of deforestation that have impoverished the country, led to unsustainable forest resource use and endangered biodiversity. To more efficiently use limited resources and prevent conflicts between sustainability targets, forest restoration should synergize climate action, life on land, and human livelihood sustainable development goals. Promoting seed dispersal by birds and mammals may achieve this by supporting diverse forests of native tree species with greater carbon storage capacity and a more diverse value for humans. In our project, we have integrated camera traps and acoustic sensors to monitor seed dispersers in restoration areas at five sites across Madagascar's humid forests. We have collected over 400 000 1-minute audio recordings from 178 ARUs. We will use OpenSoundscape to fine tune a multi-layer perceptron network on Perch embeddings to semi-automatically detect seed dispersing birds and lemurs from audio recordings. We plan to analyze the data using a multivariate GLM specifically designed for sensor monitoring to examine how distance to forest, tree species composition, vegetation structure, and reforestation age affect seed disperser communities. We are working closely with key stakeholders (reforestation NGOs in Madagascar) to develop guidelines on how to reforest to aid restoration of seed dispersers and on how to monitor this progress.