Adapter Layer

To support diverse user environments and enable interaction from both human operators and AI agents, RoboStack introduces a flexible adapter layer designed to accommodate various request types across platforms. For example, HTTP endpoints support integration with traditional services, A2A communication is optimized for AI agents, and MCP is tailored for interactions with LLMs.

This adaptable framework empowers not only human users to control robots, but also enables agents to access robot status, retrieve sensor data, and even train or control robots autonomously – laying the foundation for intelligent, real-time robotic applications.