# Peak Answer [Health: Active]

**Category:** 🎯 Marketing  
**Repository:** https://github.com/peakanswer/peak-answer-mcp  
**GitHub Stars:** 0  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/peak-answer

## Description
Whether AI search recommends your brand, which questions rivals win, and why engines skip your site.

## Tools
Capabilities this server exposes over MCP:

- **audit_site** — Answer readiness, AI crawler access, robots.txt, sitemap, structured data, headings, meta description, readability and entities for one URL, in one call. Seconds
- **generate** — Produces `llms_txt`, `faq_schema`, `content_brief`, `prompt_set` or `semantic_keywords
- **geo_audit** — Asks a real model ten real buying questions. Starts a background run and returns immediately; call again with the same domain to collect it
- **my_brand** — Which domain this connection is scoped to
- **my_visibility** — Measured visibility history for that brand, rather than a one-off check
- **my_questions** — Tracked questions, with `only_losing` for the ones a competitor wins
- **my_question_history** — One question over time
- **my_actions** — The ranked backlog for that brand
- **my_search_console** — Search Console queries near the top of page two

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What Peak Answer MCP server does

The Peak Answer MCP server provides MCP access to Peak Answer’s AI-search visibility and website analysis features. It is designed for marketing teams and developers investigating whether a brand appears in AI-generated answers, why a site may be difficult for engines to quote, and which content or technical changes deserve attention.

The server belongs in the 🎯 Marketing category because its outputs focus on brand visibility, buying questions, competitor performance, technical answer readiness, and content planning. It can work with any public URL for site audits, generation tasks, and geo audits. Brand-specific tools instead read the customer data associated with the authenticated connection.

## How it works

Peak Answer MCP server is remote-only. A client connects to `https://peakanswer.com/api/mcp`, so there is no local process or package to install. Authentication uses OAuth 2.0 with automatic discovery. Clients that manage tokens directly can use a Peak Answer API key as a bearer token.

Each connection is associated with one brand domain. Call `my_brand` to identify that domain rather than asking an agent to infer it. The tools are read-only and marked accordingly, so clients do not need confirmation before each request.

`geo_audit` starts an asynchronous evaluation using ten buying questions and returns immediately. Calling it again with the same domain retrieves the run’s results. Other tools return audits, generated outputs, or existing measured data directly.

## Setup and configuration

Add the remote endpoint to an MCP client. Claude Code can register it with:

```bash
claude mcp add --transport http peak-answer https://peakanswer.com/api/mcp
```

Clients that use `mcp.json`, including Cursor and VS Code, can define a server with the endpoint as its URL. Claude.ai and ChatGPT can add the same URL through their connector settings. An authenticated Peak Answer account is required for the MCP tools; the service is included with every Peak Answer plan.

## Tools and capabilities

The Peak Answer MCP server exposes these read-only operations:

- `audit_site` checks one URL for answer readiness, crawler access, robots.txt, sitemaps, structured data, headings, metadata, readability, and entities.
- `generate` creates an `llms_txt`, FAQ schema, content brief, prompt set, or semantic keyword output.
- `geo_audit` evaluates ten real buying questions with a model through a background run.
- `my_visibility` returns measured visibility history for the connected brand.
- `my_questions` lists tracked questions, with an option to show only questions lost to competitors.
- `my_question_history` follows one tracked question over time.
- `my_actions` returns the brand’s ranked action backlog.
- `my_search_console` finds Search Console queries close to the top of page two.
- `my_brand` identifies the domain associated with the connection.

## Limitations and notes

The service is remote-only, and its brand-scoped tools depend on the data available for the authenticated Peak Answer account. Public-URL operations can inspect public pages, but they do not represent the customer’s private measured data. The geo audit is not immediate: it begins a background run and requires a later call using the same domain to collect results. All exposed tools are read-only, so this server is intended for analysis and generation rather than making site changes.

_Full upstream README: https://allmcps.com/mcp/peak-answer/readme_