Low-code development platforms (LCDPs) are gaining momentum, since they enable less technical profiles (citizen developers) to participate in software development. LCDPs typically run on the cloud, with zero-cost installation, and often use graphical languages to describe the target applications. While building LCDPs manually is costly, the process can benefit from the automation that model-driven technologies offer.

To this end, we propose Dandelion, a cloud-based graphical-language workbench specifically designed to create LCDPs. Dandelion supports the definition of the abstract and concrete syntax of graphical languages, resulting in tailored cloud-based modelling environments. To cater for industrial needs, the tool can be connected with heterogeneous modelling technologies – including proprietary formats or standards like EMF – and can handle large models via persistence in Elasticsearch and a model pagination mechanism. To facilitate modelling by citizen developers, it features sensemaking strategies to ease understanding (meta-)models via interactive graphics and diagrams; and flexible modelling support to relax the meta-model/model conformance relation, thus tolerating inconsistencies. A video of the tool is available at https://youtu.be/xl4nzySxFvo.