aiseo-auditv2.0.1

Documentation

aiseo-audit Model Context Protocol server

In brief. This is the aiseo-audit Model Context Protocol server for giving a compatible AI assistant the same page analyzer as the command line. The audit_url tool accepts a URL, target queries, and a page profile.

Maintained by Jeff Patterson and Agency Enterprise · Updated August 30, 2026

How the aiseo-audit Model Context Protocol server works: overview

The server is the assistant-facing interface for the deterministic aiseo-audit analyzer.

A compatible client starts the published package, lists its audit_url tool, and sends a page URL with optional queries and a page profile.

The assistant receives the scored result because the server performs the measurement before the model explains it. This means model wording does not control the audit score.

aiseo-audit Model Context Protocol server terms

Model Context Protocol

Model Context Protocol refers to a standard interface between an AI client and an external tool server.

MCP server

An MCP server refers to the process that exposes aiseo-audit tools to a compatible client.

Tool call

A tool call means that the client sends structured inputs to one named server function.

audit_url

audit_url is defined as the tool that analyzes one public or local page URL.

Client

A client refers to the compatible AI application that connects to the server.

Deterministic result

A deterministic result is a type of output that stays fixed when fetched content, settings, queries, and tool version stay fixed.

Configure a client

The server runs from the published package. It needs no separate service, external AI key, or account.

{
  "mcpServers": {
    "aiseo-audit": {
      "command": "npx",
      "args": ["-y", "aiseo-audit-mcp"]
    }
  }
}

audit_url tool

Field or optionDescription
urlRequired public or local URL to analyze.
queriesOptional target-query array, up to ten.
domainOptional auto, product, or informational profile.

Interpret assistant output carefully

The server returns the same evidence-tiered result as the command line. An assistant can summarize or prioritize it. The score still measures readiness and cannot guarantee engine behavior.

How to use this reference

  1. Add the published server command to a compatible client.
  2. Restart the client and confirm that audit_url appears.
  3. Send the page URL and its target queries.
  4. Review the result before asking the assistant to draft changes.
  5. Run the same inputs again after publishing an approved edit.

Key takeaways

  • The server is a local bridge between a compatible assistant and aiseo-audit.
  • The tool is audit_url and the page URL is its only required input.
  • The query list is optional and improves page-specific alignment checks.
  • The model is the interpreter of the result, not the scorer.
  • The audit is repeatable when the fetched inputs and version stay fixed.

Bottom line: Use the server when an assistant needs to run the audit instead of guessing about page readiness.

Official sources and verification

According to the npm package page, aiseo-audit publishes its current version and installation command [1]. According to the GitHub repository, the source code and project documentation are public [2].

According to the evidence map, every scored factor records an evidence tier and pipeline stage [3]. According to the release history, major versions document scoring changes that require new baselines [4]. According to the project license, aiseo-audit uses the MIT license[5].

  1. aiseo-audit on npm: package, version, and installation details.
  2. agencyenterprise/aiseo-audit: source code and documentation.
  3. aiseo-audit evidence map: factor tiers, stages, and research sources.
  4. aiseo-audit releases: version history and migration notes.
  5. MIT license: project license text.