AEO NIUEXA

ChatGPT, Claude and AI Agents (MCP)

6 min read

Use AEO NIUEXA From Your AI Assistant

AEO NIUEXA has an MCP (Model Context Protocol) server, so AI assistants such as ChatGPT and Claude, and coding agents such as Claude Code, can run scans, read results and fixes, and work with AI Visibility on your account. The server speaks Streamable HTTP at https://aeo.niuexa.ai/mcp/mcp. Scans and AI Visibility checks use the credits in your account, at the same prices as in the app.

Connect ChatGPT or Claude (Sign-In)

No API key is needed: you sign in to your AEO NIUEXA account and approve access.

  1. In ChatGPT or Claude, open the apps or connectors settings and add a custom connector with the URL https://aeo.niuexa.ai/mcp/mcp (or add AEO NIUEXA from the app's directory where it is listed)
  2. Sign in to AEO NIUEXA in the window that opens and choose Allow
  3. Ask, for example: "Scan example.com for AI readiness and list the most important fixes" or "How visible is my brand in ChatGPT and Perplexity answers?"

Your email address must be confirmed before you can approve a connection. Connected apps are listed under Profile → Connected apps, where you can disconnect them.

Connect With an API Key

For coding agents and clients that take a token, create an API key and send it in the Authorization header as Bearer YOUR_API_KEY.

  1. Open the account menu and choose API keys
  2. Name the key after the device or tool that will use it (when you signed in a while ago, enter your password too) and choose Generate key
  3. Copy and save the key: it is shown only once
  4. Add the server to your agent's configuration
The API keys page: create a named key, and a ready-to-copy MCP client configuration with the server address and the Authorization header.
API keys: one named key per device or tool, each revocable on its own.

Claude Code:

claude mcp add --transport http aeo-analyzer https://aeo.niuexa.ai/mcp/mcp --header "Authorization: Bearer YOUR_API_KEY"

Clients configured with JSON:

{
  "mcpServers": {
    "aeo-analyzer": {
      "type": "http",
      "url": "https://aeo.niuexa.ai/mcp/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Keys act as your account: treat them like passwords, use one per client and revoke a key at once if it may have been exposed. Keys can be created only after your email is confirmed.

Available MCP Tools

Analysis tools start with aeo_; AI Visibility and content tools start with mt_. Tools that cost credits say so; the others are free (some have daily limits).

Tool Description
aeo_start_scanStart a scan of one page (costs 1 credit)
aeo_get_scan_statusCheck a scan's progress
aeo_list_scansList recent scans
aeo_get_scan_resultsGet scores, checks and summary data
aeo_get_failed_checksGet the failed checks with their advice, most severe first
aeo_get_check_detailsGet one check with the analyzer's raw data
aeo_get_recommendationsGet the ranked fixes for one analysis
aeo_get_creditsCheck your credit balance and usage
mt_list_brandsList your tracked brands with summary figures
mt_get_brand_summaryGet one brand's summary: visibility, sentiment, citations and category scores
mt_get_visibility_trendsGet a brand's daily visibility over a period
mt_list_mentionsList and filter a brand's stored answers
mt_get_mentionGet one answer with its question, text and citations
mt_list_citationsList the pages AI engines cited, yours and others'
mt_list_content_gapsList questions where others are cited and you are not
mt_list_briefsList a brand's content briefs
mt_get_briefGet a complete brief
mt_list_articlesList a brand's articles
mt_get_articleGet an article in Markdown with its metadata
mt_list_alertsList alerts for your brands
mt_create_brandCreate a tracked brand for one of your sites
mt_trigger_checkRun an AI Visibility check now (costs 10 credits); weekly monitoring is set up on the brand page
mt_generate_briefGenerate a brief from a content gap (costs 1 credit)
mt_generate_articleGenerate an article from a brief (costs 2 credits)
mt_generate_social_postsGenerate social posts from an article (costs 1 credit)
mt_update_articleUpdate an article's title, text or status
mt_list_promptsList a brand's questions and which are tracked
mt_get_prompt_performanceGet each question's results across past checks
mt_generate_promptsSuggest questions from the brand's site and business description (free, daily limit)
mt_add_promptAdd your own question
mt_update_promptEdit a question or choose whether it is tracked (at most 10)
mt_delete_promptArchive a question; its past answers are kept
mt_update_business_contextEdit the business description questions are written from
mt_list_factsList a brand's approved facts and the wrong facts AI answers gave
mt_manage_factAdd, edit or archive an approved fact

What the Tools Return

  • Results follow the same rules as the reports: checks that do not apply have the status not_applicable and never count as passed, results judged by a retired rule are left out, and unproven checks are only counted.
  • Fixes are ranked by severity, then by how much the check counts, each with an impact word, its stage, rule ids and evidence level. No fix comes with a promised number of points.
  • Two resources describe the scoring: aeo://scoring-model (dimension weights, grades and evidence multipliers) and aeo://check-reference (every check with its evidence level). Both are built from the live scorer, so they cannot drift from it.