Roba Labs schedules Copilot v2 for June
- Roba Labs will launch Roba Copilot v2 on June 1 — an AI-native, no-code robotics studio for building, editing and simulating robots from natural-language descriptions. - Roba Copilot claims users can describe robots and tasks without ROS or coding, positioning itself as a 'Cursor for robotics' for rapid prototyping by non-programmers. - The move targets democratizing robot workflows but raises questions about integration with existing stacks and safety checks. (x.com)
Roba Labs is teeing up a June 1 release for Copilot v2, and the pitch is unusually direct: describe a robot and a task in plain language, then build, edit and simulate it without touching ROS or writing code. That launch framing comes from a social post citing Roba Labs, while Roba’s own site and prior launch materials describe a broader “people’s robotics platform” built around no-code tooling, browser-based simulation and an open, interoperable workflow. (robalabs.com) Why that matters: most robotics workflows still assume users can navigate middleware, simulators, packages, hardware interfaces and deployment steps across multiple tools. Roba’s earlier public materials said the company wanted to replace “10+ fragmented platforms” with a single workflow spanning build, train, share and monetize, and described ROBA Link as a layer that connects third-party tools, SDKs and APIs. Copilot v2 looks like the next step in that same product direction: moving the interface for robotics creation closer to chat-first software tooling. (thetradable.com) The company has been building toward this for months. In October 2025, Roba Labs launched the first public version of its ecosystem and called it “The Hugging Face of Robotics.” That announcement emphasized open hardware and software, a physics simulator, and user ownership of data, assets and revenue. A company profile published later described Roba Studio as browser-accessible and “ROS/Zenoh-native,” with templates for pick-and-place, Nav2 navigation, SLAM, perception, calibration and reinforcement learning agents. (thetradable.com) So the core claim behind Copilot v2 is not just “AI for robots.” It is that robotics authoring can be abstracted upward: from code and middleware into intent, prompts and generated workflows. In software, that is the comparison behind the “Cursor for robotics” line. In practice, the test will be whether Roba can reliably translate natural-language requests into robot structures, task logic and simulation setups that users can inspect and modify rather than merely demo. That last point is an inference from the product claims and the company’s existing emphasis on simulation and editing tools. (thetradable.com) There are two obvious constraints. First, integration: Roba’s own materials say the platform connects outside tools and APIs, which suggests it still has to coexist with established robotics stacks rather than replace them outright. Second, safety: Roba previously used the phrase “learn fast in simulation, ship safely in reality,” but public materials reviewed here do not spell out what validation, guardrails or approval steps Copilot v2 will require before anything moves from prompt to physical deployment. (thetradable.com) The near-term question is simpler than the long-term ambition. On June 1, users will be able to judge whether Copilot v2 is a real no-code robotics studio with editable outputs and usable simulation, or mainly a polished front end for ideas. The company’s website, GitHub research hub and launch materials show Roba has already assembled pieces in simulation, browser tooling and embodied-AI research; Copilot v2 is the clearest public test yet of whether those pieces can be packaged for non-programmers. (robalabs.com)