EdgeAI Forge uses local AI agents to generate ROS 2 vision pipelines with human review
This local-first tool turns plain-English robotics requirements into ROS 2 package designs, but halts before any code runs.
EdgeAI Forge is a local-first engineering architecture that aims to generate, test, benchmark, and document ROS 2 and machine-vision pipelines. The current proof of concept focuses on a ROS 2 vision pipeline generator: given a requirement like reading frames from a USB camera and running object detection, it produces design artifacts including package manifests, Python or C++ nodes, topic/message definitions, parameters, launch files, tests, container packaging, and performance benchmarks. The system uses three prompt agents — Planner, Vision, and ROS — backed by a local Ollama endpoint, meaning no cloud AI is required. The broader scaffold adds API, dashboard, infrastructure, and observability components, though generated output today is design-level, not production-ready.
The project's central safety boundary is that generated robot-motion or production-deployment artifacts should never run automatically. The intended workflow includes simulation, test evidence, human review, explicit approval, logging, and rollback. The creator is actively seeking community feedback on five key questions: which package archetype should be the first template, how generated packages should be evaluated beyond compilation, which simulation patterns (launch_testing, rosbag, Gazebo) should be mandatory, how to encode QoS, lifecycle nodes, diagnostics, and hardware assumptions, and what safeguards prevent unsafe deployment practices. This positions EdgeAI Forge as a thoughtful alternative to earlier tools like ROScribe, with a stronger emphasis on human oversight in safety-critical robotics.
- Uses three Ollama-backed agents (Planner, Vision, ROS) to convert natural-language requirements into ROS 2 design artifacts.
- Generates package manifests, Python/C++ nodes, launch files, tests, Docker/Jetson packaging, and FPS/latency benchmarks.
- Enforces human-in-the-loop safety via simulation, test evidence, explicit approval, logging, and rollback before deployment.
Why It Matters
For robotics teams, EdgeAI Forge could slash ROS 2 development time while keeping humans in control of safety-critical deployments.