The latest release of dora-rs/dora is v1.1.0. DORA is middleware for running robot and AI applications as dataflows, and the name stands for Agentic Dataflow-Oriented Robotic Architecture. An application is modeled as a directed graph, also called a pipeline. The release notes for v1.1.0 are, however, not part of the research material, so what changed in this version cannot be confirmed.
The project describes itself in its README as a framework built 100% in Rust. It ships YAML dataflow files and a single CLI, and it can combine Rust, Python, C, and C++ nodes inside one dataflow. It connects robot nodes written in different languages and manages the lifecycle of local and distributed runs, and it also supports monitoring, record and replay, and OpenTelemetry-based logging, metrics, and distributed tracing. The license is Apache-2.0, and the repository has 4028 stars.
The evidence stops there. Real operational cases and performance measurements are not in the research material. It is not known at what scale the low-latency, distributed dataflow claimed by the repository description has been validated, and the evaluation score stops at 3.4 out of 10 on the momentum item. The freshness, technology, and quality items are empty. Whether to use this middleware for a robot application is better decided after seeing the next release and more real deployments.