Skip to content

Rootpublish MCP server

Which MCP servers are Rootpublish's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-25

← Back to Rootpublish

Whose server it is
  • Registry namecom.rootpublish/ai-factsverified in the MCP registry

The registry name sits under rootpublish.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-10-25
  • Stateanswerschecked 2026-10-25
  • Tools3
  • Server calls itselfrootpublish-ai-facts-remote
  • Server version0.1.0
  • Protocol version2025-06-18

On 2026-10-25 the server answered initialize and tools/list without a sign-in. First answer in 814 ms.

Should an agent connect itNot proven either way yet
  • Whether its tools do the product's job has not been judged yet.
  • Checked on 3 of the 7 days it takes to judge reliability.

Once connected

  • For "find wrong plans or contract terms in AI answers", the likely first call is buyer_questions.
  • Parameters are untyped or undescribed in 2 of 3 tools: check arguments before calling.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 3, the first on 2026-10-25
  • INITE MCP scoreafter a weeka score needs 7 daily checks: 3 so far
  • Answeredmeasuring3 of the 7 daily checks it takes to print a share
  • Median first answer817 ms
  • Space in an agent's context≈ 433 tokensthe definitions of all its tools (3), which an agent carries on every turn
  • Tool definitionsC · 72 of 100
  • Typed results0 of 3 toolsdeclared outputSchema: results come typed, read once
  • Parameters without a description: in 2 of 3
  • Not saying whether it only reads or can destroy: 3 of 3

What an agent can do with it

  • not sure · check what AI assistants say about your prices and terms
  • can do · find wrong plans or contract terms in AI answers
  • not sure · track how AI answers change across repeated buyer questions

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

None of the five warning signs in these definitions.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers3 tools · read 2026-10-25
  • buyer_questions

    Start an AI facts check: returns the questions a buyer asks about a company. Answer them yourself, with web search, before reading anything from the company.

  • company_statements

    After you have answered the buyer questions: reads the company's pricing and service pages and returns its statements with instructions for comparing your answers.

  • verify_findings

    Keeps only the differences whose quotes are really in your answer and on the company's page, and finds the pages your answer cited that carry a wrong figure. Call once per answer.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "ai-facts": {
      "type": "http",
      "url": "https://rootpublish.com/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

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

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://rootpublish.com/mcp as a custom connector in the settings.

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.