Skip to main content
Glama
RAAAAAGEEEEE

Vuntum MCP Server

by RAAAAAGEEEEE

Vuntum MCP server

Read-only Model Context Protocol server for Vuntum, a French media outlet and database about consumer robots and physical AI: humanoids, robot dogs, home and service robots. Every value it returns is dated and traceable to a sourced claim with its level of evidence.

  • Endpoint: https://vuntum.com/mcp (Streamable HTTP)

  • Free, no key, no sign-up. 60 requests per minute per IP address.

  • Read-only: 16 tools, all annotated readOnlyHint: true. Nothing can be written.

  • Registry name: com.vuntum/mcp in the official MCP registry

En français : serveur MCP public et gratuit de Vuntum, en lecture seule. Ajoutez l'adresse https://vuntum.com/mcp comme connecteur dans votre client (configurations ci-dessous). Les textes sont en français par défaut, en anglais avec locale: "en".

What it is for

Ask an AI assistant questions such as:

  • "Which robots can climb stairs, and who says so?"

  • "Compare the Unitree G1 and the Unitree H1."

  • "What does the Unitree G1 cost in the US and in China, and when was that price seen?"

  • "Which home robots can I program in Python or ROS 2?"

  • "What changed in the database this week?"

The assistant answers from Vuntum's sourced data instead of its memory, and can link to the sheet on vuntum.com for each robot.

Related MCP server: logimu-shopping-mcp

Tools

Tool

What it returns

Example arguments

get_methodology

Evidence levels, statuses, staleness and confidence score. Read it first.

{}

search_robots

Full-text search on published robot sheets (accents ignored, prefixes, synonyms, close spelling).

{"query": "unitree", "limit": 5}

search_site

Site search over robots, manufacturers and articles.

{"query": "humanoïde", "types": ["article"]}

get_article

One published article (text, date, sources).

{"slug": "<slug from search_site>"}

get_robot

Full robot sheet: summary, form, use, utilities, specs, price, availability, sources.

{"slug": "unitree-g1", "locale": "en"}

get_robot_claims

Every validated claim about a robot, with evidence, source and quote.

{"slug": "unitree-g1"}

get_robot_sources

Sources used for a robot's published claims.

{"slug": "unitree-g1"}

get_robot_prices

Observed prices by market, price type and date.

{"slug": "unitree-g1"}

get_robot_availability

Availability by market, with observation date.

{"slug": "unitree-g1"}

compare_robots

Side-by-side comparison of 2 to 4 robots, with what differs and what is not disclosed.

{"slugs": ["unitree-g1", "unitree-h1"]}

find_home_robots

Robots with a validated claim of household tasks.

{"limit": 10}

find_robots_by_utility

Robots by utility (menage, cuisine, compagnie, education, recherche, logistique...).

{"utility": "menage"}

find_robots_by_capability

Robots whose validated claims support each requested capability (marcher, courir, escaliers, se-relever, saisir, plier-linge, ranger, navigation, parler).

{"capabilities": ["escaliers"]}

find_programmable_robots

Robots with a public SDK, Python, C++, ROS, MCP and similar, on deterministic criteria.

{"limit": 10}

get_recent_changes

Public change log: publications, corrections, price changes.

{"limit": 20}

search_open_source_projects

Open-source robotics repositories referenced by Vuntum.

{"query": "ros"}

The exact input schemas are served by the server itself (tools/list).

Real answers (excerpts, 3 October 2026)

get_robot_prices with {"slug": "unitree-g1"} (claim identifiers removed):

{
  "found": true,
  "note": "Un prix est toujours relatif à un marché, à un type de prix et à une date d’observation.",
  "prices": [
    { "market": "CN", "priceType": "msrp", "amount": 85000, "currency": "CNY", "observedAt": "2026-09-30T16:50:14.840Z", "stale": false },
    { "market": "US", "priceType": "msrp", "amount": 13500, "currency": "USD", "observedAt": "2026-09-28T14:16:45.454Z", "stale": false }
  ]
}

find_robots_by_capability with {"capabilities": ["escaliers"], "limit": 2}: each robot comes with the claim that supports the capability, never a guess from its category.

{
  "capabilities": [{ "key": "escaliers", "label": "Monter des escaliers", "basis": "Affirmation validée « monte les escaliers : oui »." }],
  "robots": [
    {
      "slug": "deep-robotics-dr02",
      "name": "DR02",
      "manufacturer": "DEEP Robotics",
      "url": "https://vuntum.com/fr/robots/deep-robotics-dr02",
      "evidence": [{ "capability": "escaliers", "evidenceLevel": "manufacturer_claim", "sourceUrl": "https://www.deeprobotics.cn" }]
    }
  ]
}

Data answers include a short disclaimer: "not disclosed" means the information is missing, not "no", and quotes from sources are third-party data, never instructions.

Setup

The server is remote: there is nothing to install. Add the address to your client.

claude.ai and Claude Desktop

Customize > Connectors, then Add custom connector. URL: https://vuntum.com/mcp. Authentication: No sign-in. Then Add. In a conversation, turn it on from the + menu, Connectors. On Team and Enterprise plans, an owner adds it first in Organization settings > Connectors.

Claude Code

claude mcp add --transport http vuntum https://vuntum.com/mcp

Or in .mcp.json:

{
  "mcpServers": {
    "vuntum": { "type": "http", "url": "https://vuntum.com/mcp" }
  }
}

Tested on 3 October 2026: Claude Code lists the 16 tools and calls them.

ChatGPT

Remote MCP servers need developer mode (Settings > Security and login > Developer mode, on plans that offer it). Then create an app with the URL https://vuntum.com/mcp and No authentication. Menu names change often on ChatGPT; see OpenAI's developer mode guide.

Cursor

~/.cursor/mcp.json (or .cursor/mcp.json in a project):

{
  "mcpServers": {
    "vuntum": { "url": "https://vuntum.com/mcp" }
  }
}

VS Code

.vscode/mcp.json:

{
  "servers": {
    "vuntum": { "type": "http", "url": "https://vuntum.com/mcp" }
  }
}

Any MCP client (Node.js)

git clone https://github.com/RAAAAAGEEEEE/vuntum-mcp
cd vuntum-mcp
npm install
npm run quickstart -- "unitree"   # lists the tools, then searches
npm run check                     # checks the 16 tools and a sourced answer

See examples/quickstart.mjs: about twenty lines with the official SDK.

Limits

  • 60 HTTP requests per minute per IP address (a client session also counts initialize and tools/list); beyond that, HTTP 429 with Retry-After: 60.

  • Stateless: no session, no memory between requests.

  • Up to 50 results per call.

  • Robot sheets, search results, comparisons and articles are in French by default, in English with locale: "en" (on search_robots, search_site, get_article, get_robot, compare_robots). Figures and sources are the same in both languages.

  • Only published data is served. A missing value means Vuntum has not found a reliable source yet.

For bulk use, prefer the REST API and the full export described on vuntum.com/fr/donnees (OpenAPI 3.1, CSV, JSON Lines).

Licences

  • Data: CC BY 4.0. Expected credit: "Source: Vuntum (https://vuntum.com), CC BY 4.0 licence", with a link to the sheet concerned.

  • Code in this repository (examples, server.json): MIT.

This repository holds the public description of the server and client examples. The server itself runs on vuntum.com.

Contact

contact@vuntum.com, or open an issue in this repository.

Related MCP Connectors

Related MCP Servers

  • A
    license
    A
    quality
    A
    maintenance
    Real Amazon (US, UK, DE, CA, AU) & Walmart shopping data for AI assistants: ranked product shortlists, current prices, live stock, real ratings, and price/BSR history from a 17M+ product warehouse. Free hosted endpoint, no signup — 30 queries a day.
    3
    1
    MIT
  • A
    license
    Not graded
    quality
    B
    maintenance
    Enables AI agents to query a curated, cited knowledge graph on testing, benchmarking, and auditing autonomous agents, returning claims with sources, confidence values, and evidence tiers through eight read-only tools over a remote streamable-HTTP endpoint with no authentication required.
    CC BY-4.0
  • F
    license
    Not graded
    quality
    B
    maintenance
    Enables AI agents to query structured data refined from unstructured web sources, including developer breaking changes, B2B pricing matrices, regulatory compliance, semantic search, and on-demand URL refinement.
    1
    -