About

About AT-SYNAPSE

AT-SYNAPSE connects MCP-compatible language-model agents with ROS 2 robot functions. Voice or text instructions can be grounded in camera and 3D-LiDAR inputs, then routed to vision, speech, face, motion, navigation and safety tools. A browser console using rosbridge and WebSocket exposes robot state and the agent’s reasoning process to operators.

AttracLab describes support across AMRs, UGVs, UAVs and USVs as a software integration scope rather than a single robot morphology. A later exact-product demonstration used AT-SYNAPSE on a VLM-equipped rail robot for repeated site traversal and railway inspection; that configuration supports demonstrated inspection maturity but does not imply the same task is turnkey on every supported platform.

Timeline

AttracLab develops VLM rail robot for repeated site inspection
2026-08-13 - Capability - The robot travels on commercially available 30 mm square aluminium rails and supports LTE or Wi-Fi connectivity with browser-based remote monitoring. AT-SYNAPSE aligns observations from repeated patrols, while the vision-language model identifies equipment and site changes and describes them in natural language. AttracLab plans field pilots and further validation following the development release.
Source
AttracLab launches AT-SYNAPSE physical-AI platform for autonomous vehicles
2026-08-04 - Product Launch - AT-SYNAPSE combines higher-level decision logic with ROS-based vehicle control and fleet monitoring. The platform is designed to work across UAV, UGV and USV form factors through common interfaces. Its decision layer predicts multiple possible future actions before selecting and executing a vehicle response.
Source
AttracLab adds real-time reasoning console to AT-SYNAPSE
2026-07-24 - Capability - The Web Console displays camera feeds, battery voltage, wheel speed, obstacle data and time-stamped AI reasoning in one browser view. Brain on Desktop lets an MCP-compatible agent use ROS 2 robot tools from an external computer. The interface also supports NVIDIA Isaac Sim workflows for evaluating perception, navigation and safety before hardware deployment.
Source
AttracLab launches AT-SYNAPSE for MCP-controlled ROS 2 robots
2026-07-23 - Product Launch - The platform accepts voice or text instructions and combines camera, 3D LiDAR and other sensor inputs with an AI-agent control engine. Six reusable tool groups cover vision, speech, facial display, motion, navigation and safety monitoring, with a web console showing robot state and reasoning flow.
Source