md is a declaration layer and tooling suite designed to connect robots with Anthropic Claude and other Model Context Protocol (MCP)-aware agents, enabling compliant, auditable, and deterministic control of robotic systems. It focuses on generating signed audit bundles for every robotic action, ensuring traceability and regulatory compliance, particularly with the EU AI Act §§22–26. The tool is intended for developers and operators who need to wire robots to AI planning agents while maintaining strict safety, transparency, and auditability standards.
The platform operates by authoring a single manifest file per robot, which declares the robot's identity, capabilities, and safety constraints. This file contains YAML frontmatter that is trusted by validators and a prose body readable by planning agents. md integrates directly with Claude Code and supports other MCP-compatible surfaces such as Gemini CLI and OpenAI Codex, with pre-release support for ChatGPT and Amazon Q. The tool can be installed as a Claude plugin or via Python's pipx, and it provides both manifest authoring for planning and, when paired with the robot-md-gateway, enforcement of actions on physical hardware. Safety is enforced at the gateway layer, ensuring that declarations alone do not result in robot movement.
md is its compliance moat, which includes the automatic generation and submission of five types of RCAN-signed attestation packets to the Robot Registry Foundation (RRF). These cover fundamental rights impact assessments, safety benchmarks, instructions for use, structured incident reporting, and EU register filings, all linked to the robot's registry record. Each attestation is signed with the operator's release key and stored in a transparency log, supporting regulatory and audit requirements.
md serves as the reference declaration layer for the RCAN protocol, an open standard for signed robot communications authored by the tool's creators. It is designed for zero runtime lock-in and operates across multiple AI agent platforms and infrastructures. The tool is positioned as a foundational component within the OpenCastor robotics stack and is registry-resolvable via the Robot Registry Foundation.
ROBOT.md is an Autonomous agents & workflows product. It focuses on automating robot control and generating verifiable audit trails using Claude and MCP. It is built as an open-source project for developers building AI-powered robotics workflows. ROBOT.md is open source under the Apache-2.0 license. The product ships for the command line and API, and it can be self-hosted.
ROBOT.md first shipped in 2026. Development happens publicly on GitHub with 643 commits in the last 90 days. PulseGate's similarity index finds few close equivalents — ROBOT.md occupies a relatively distinct niche. Key capabilities include robot integration, audit bundle generation, and MCP native. It exposes integrations via an MCP server.
Latest indexed changes and source events
robotmd.dev discovered by the PulseGate indexer
Other apps tracked under the same category.
