What did Hlido score RobotMCP?

RobotMCP scored 81/100 (STEADY) in Hlido's independent, hands-on review.

How much does RobotMCP cost?

Hlido recorded the pricing model as: Open source. Derived from evidence Hlido holds, not vendor-supplied; verify current prices on the vendor's pricing page.

Does any vendor pay Hlido for placement?

No. Hlido takes no money from the agents it rates — scoring weights stay private and the evidence behind every verdict is public.

RobotMCP

Name: RobotMCP review
Item: RobotMCP
Rating: 81
Author: Hlido Editor

Specialized verticals · tested 2026-08-27 · by the Hlido desk, not the vendor

In short: One MCP server that turns natural language into ROS topics, services and actions across drones, quadrupeds, mobile manipulators and industrial arms — Apache 2.0, in the official MCP Registry, with no changes to your robot code.

4 PASS · 1 FAIL of 5 public-surface claims

Visit RobotMCP → scorecard.json

Quick answer

RobotMCP scores 81/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-08-27). STEADY (81) because the portability claim is demonstrated across four genuinely different machine classes under three different models rather than asserted, adoption and provenance are strong (1,200+ stars, official MCP Pricing: Open source (free entry point documented).

Hlido Editor · 2026-08-27

RobotMCP is the most convincing demonstration in this batch of a genuinely hard integration problem being reduced to one connector. It bridges the Model Context Protocol to ROS 1 and ROS 2, so an instruction typed to Claude, ChatGPT or Gemini becomes ROS topic, service and action calls. The claim that earns the score is breadth demonstrated rather than asserted: four classes of machine — an aerial drone via Gemini, a legged robot via Cursor, a mobile manipulator via Gemini and a warehouse unloading arm via Claude — each shown running the same MCP server with a different model and no per-robot integration work. Multi-robot, multi-model portability is exactly the claim most robotics middleware fails, and showing four distinct machines under different models is a much stronger form of evidence than a feature matrix. The supporting signals are consistent: 1,200+ GitHub stars, officially listed in the MCP Registry, Apache 2.0, installable with a single pipx command, and stated to require no changes to robot code beyond adding a rosbridge node. It names work with KUKA, Doosan, Intrinsic (Google) and the Linux Foundation, which for an industrial-robotics connector is a meaningful credibility signal. What we cannot verify from a public surface, and what a buyer must weigh heavily here, is safety. This is a natural-language actuator for physical machines that can injure people. The reviewed surface describes installation, portability and a Cloudflare tunnel to your robot, but does not document safety interlocks, command validation, e-stop integration, or what happens when a model misinterprets an instruction — and the tunnel itself deserves scrutiny before it is pointed at industrial hardware. The demo clips are vendor-selected successes; failure behaviour is not shown.

Why STEADY

STEADY (81) because the portability claim is demonstrated across four genuinely different machine classes under three different models rather than asserted, adoption and provenance are strong (1,200+ stars, official MCP Registry listing, Apache 2.0, named work with KUKA, Doosan, Intrinsic and the Linux Foundation), and installation is a single command with no robot-code changes. Held below VITAL because this is a natural-language actuator for physical machines and the public surface documents no safety interlocks, command validation or e-stop integration — the demos show selected successes, never a misinterpreted instruction.

Public-surface checklist

PASS Homepage loads (required)
PASS Primary value prop (required) — 'Any robot. Any AI. A single connector.'
PASS Cta present (required) — 'Get Started in 5 Minutes' / 'Star on GitHub'
PASS Pricing or access (required) — Open source and free under Apache 2.0; single pipx install command published, no paid tier advertised
FAIL Safety posture stated (required) — No safety interlocks, command validation or e-stop integration documented for a natural-language physical actuator

What we saw

4 screenshots captured by the Hlido engine during the reviewed run (run-c2ea691ab0bf9fee-robotmcp-ai). Our own captures — not vendor marketing material.

RobotMCP — run screenshot 1 (home.png) — home.png

RobotMCP — run screenshot 2 (page____docusaurus_skipToContent_fallback.png) — page____docusaurus_skipToContent_fallback.png

RobotMCP — run screenshot 3 (page_.png) — page_.png

RobotMCP — run screenshot 4 (page_intro.png) — page_intro.png

What it does well

Demonstrates the same MCP server driving four different machine classes — aerial, legged, mobile manipulation and industrial — under three different models
Model-agnostic and robot-agnostic: works with Claude, ChatGPT, Gemini or any MCP client, across ROS 1 and ROS 2
No changes to robot code required — add a rosbridge node and connect
Apache 2.0 and officially listed in the MCP Registry, with 1,200+ GitHub stars
Single-command install via pipx with modules downloaded on first run
Names collaboration with KUKA, Doosan, Intrinsic (Google) and the Linux Foundation
Translates instructions into ROS topics, services and actions rather than a proprietary abstraction

What it fails at

No safety interlocks, command validation or e-stop integration documented on the reviewed surface — the critical gap for a physical actuator
Behaviour when a model misinterprets an instruction is not shown or described
The Cloudflare tunnel to your robot is presented as a convenience without a security discussion
Every demo clip is a vendor-selected success; no failure case is published
No pricing, commercial support or SLA statement for what would be industrial deployment
Natural-language control of machinery that can injure people demands assurance the public surface does not provide

Best for

Robotics researchers and ROS developers prototyping natural-language control across heterogeneous fleets
Teams evaluating multi-robot, multi-model portability without per-robot integration work
Lab and simulator environments where a misinterpreted instruction is not dangerous

Not recommended for

Production industrial deployment without an independent safety review — the surface documents no interlocks
Any setting where a misinterpreted natural-language command could injure a person or damage equipment
Teams requiring commercial support or an SLA for safety-critical machinery

Pricing & access

ModelOpen source
Free entry pointYes — a free tier or open-source edition is documented
Pricing findable on the public surfacePASS Open source and free under Apache 2.0; single pipx install command published, no paid tier advertised (tested 2026-08-27)

Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-08-27.

Compared to

DroidMind ros-robot-fleet-reach-vs-documented-confirmation-gate
Both expose physical-device control to an AI assistant over MCP — DroidMind to Android handsets, RobotMCP to ROS robots. DroidMind explicitly states a confirmation gate on high-risk actions; RobotMCP demonstrates far broader hardware reach but documents no equivalent safety gate. Different blast radius, and RobotMCP’s is larger.

Agent relevance

CLI MCP Behavioral-testable

Agentic-Commerce Readiness 64/100 · INTEGRABLE

Independent readiness for agent delegation & transaction. How it’s scored · check live

Install via pipx, add a rosbridge node to the robot, start the server (which opens a Cloudflare tunnel), then point Claude, ChatGPT, Gemini or any MCP client at it. The agent issues natural-language instructions that the server translates into ROS topics, services and actions. This is a high-consequence physical actuator, not a read-only tool — treat granting it to an autonomous agent as a safety decision, not a configuration one.

Agent-friendly score: 9/10

Evidence

One MCP server drives four robot classes under three different models with no per-robot integration work — source (2026-08-27) verified
Natural-language instructions become ROS topics, services and actions across ROS 1 and ROS 2, with any MCP client — source (2026-08-27) verified
Apache 2.0, 1,200+ GitHub stars, officially in the MCP Registry; built with KUKA, Doosan, Intrinsic (Google) and the Linux Foundation — source (2026-08-27) verified
Single-command install with no changes to robot code; opens a Cloudflare tunnel to the robot — source (2026-08-27) verified

scorecard.json · transparency passport · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-1+editorial-narrative-v2 · Methodology version 2026.05 ·

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/lpigeon-ros-mcp-server.svg)](https://hlido.eu/check/?agent=lpigeon-ros-mcp-server)

HTML

<a href="https://hlido.eu/check/?agent=lpigeon-ros-mcp-server"><img src="https://hlido.eu/badge/lpigeon-ros-mcp-server.svg" alt="Hlido trust score"></a>