Dora is a dataflow framework, written entirely in Rust, for building real-time robotics and embodied-AI applications. Short for Agentic Dataflow-Oriented Robotic Architecture, it models an application as a pipeline of nodes that exchange data with very low latency.
In Dora, applications are directed graphs: camera drivers, perception models, planners, and controllers are separate nodes connected into a dataflow, which makes systems composable and easy to distribute across machines. Communication uses shared memory where possible to keep latency low, a critical property for robots that must react in real time. The framework offers both Rust and Python APIs, a command-line tool for running and managing dataflows, and a growing ecosystem of ready-made nodes for hardware and AI models. Dora reached its 1.0 release as a mature platform, is open source under the Apache-2.0 license, and has over 4,000 stars on GitHub, with active development continuing through 2026.
Dora is free and open source under the Apache-2.0 license. There is no paid tier; you run it on your own robots and computers.
Robotics engineers and embodied-AI developers who need a fast, composable way to wire sensors, models, and actuators into real-time systems.
Dora positions itself as a modern alternative to robotics middleware stacks that are harder to compose across languages. Because nodes communicate through a simple dataflow description, teams can prototype in Python and later move performance-critical nodes to Rust without redesigning the system, and dataflows can be recorded and replayed, which makes debugging real robot behavior considerably easier.