<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: ReputeMap</title>
    <description>The latest articles on DEV Community by ReputeMap (@reputemap).</description>
    <link>https://dev.to/reputemap</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4245908%2Ffb53fe30-8914-494d-9d05-225bfb41579b.png</url>
      <title>DEV Community: ReputeMap</title>
      <link>https://dev.to/reputemap</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/reputemap"/>
    <language>en</language>
    <item>
      <title>Ask your Google reviews from Claude: a hosted MCP server walkthrough</title>
      <dc:creator>ReputeMap</dc:creator>
      <pubDate>Sun, 15 Nov 2026 12:19:54 +0000</pubDate>
      <link>https://dev.to/reputemap/ask-your-google-reviews-from-claude-a-hosted-mcp-server-walkthrough-2bbp</link>
      <guid>https://dev.to/reputemap/ask-your-google-reviews-from-claude-a-hosted-mcp-server-walkthrough-2bbp</guid>
      <description>&lt;p&gt;If you manage Google reviews for more than one location, the day usually starts the same way: open the Business Profile for each location, scroll, find what's new, find what's still unanswered. I wanted to ask one question instead: &lt;em&gt;"which of my locations has the most unanswered reviews, and what are people complaining about?"&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;That's what the Model Context Protocol (MCP) is good at. Here's how we wired Google reviews into Claude, Claude Code and Cursor with a hosted MCP server, and the design decisions that mattered.&lt;/p&gt;

&lt;h2&gt;
  
  
  The setup in one command
&lt;/h2&gt;

&lt;p&gt;For Claude Code:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;claude mcp add --transport http reputemap https://api.reputemap.com/mcp \
  --header &lt;span class="s2"&gt;"Authorization: Bearer rmk_YOUR_KEY"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Then:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;claude &lt;span class="s2"&gt;"which of my locations has the most unanswered reviews?"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;For Cursor, add this to &lt;code&gt;~/.cursor/mcp.json&lt;/code&gt;:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;
  &lt;span class="nl"&gt;"mcpServers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
    &lt;span class="nl"&gt;"reputemap"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
      &lt;span class="nl"&gt;"url"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"https://api.reputemap.com/mcp"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
      &lt;span class="nl"&gt;"headers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="nl"&gt;"Authorization"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"Bearer rmk_YOUR_KEY"&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;
    &lt;span class="p"&gt;}&lt;/span&gt;
  &lt;span class="p"&gt;}&lt;/span&gt;
&lt;span class="p"&gt;}&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;In claude.ai or Claude Desktop you don't need a key at all: add it as a connector and sign in through OAuth.&lt;/p&gt;

&lt;h2&gt;
  
  
  The tools
&lt;/h2&gt;

&lt;p&gt;The server is Streamable HTTP and exposes seven tools:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Tool&lt;/th&gt;
&lt;th&gt;What it does&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;list_locations&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Every location you manage, with ids&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;list_reviews&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Reviews filtered by location, rating, unanswered, last N days, with the published reply&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;get_review_stats&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Counts, averages, 1–5 breakdown, unanswered, per location and in total&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;draft_reply&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Reply variants, saved as a draft. Never publishes&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;publish_reply&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Publishes a reply you explicitly approved&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;send_review_request&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Sends a real review-request email&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;get_usage&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;This month's API credits&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Three design decisions worth copying
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;1. Read tools run freely, write tools always ask.&lt;/strong&gt; Every tool carries MCP annotations: the four read tools are marked read-only, while &lt;code&gt;draft_reply&lt;/code&gt;, &lt;code&gt;send_review_request&lt;/code&gt; and &lt;code&gt;publish_reply&lt;/code&gt; are not. Well-behaved clients ask before running them, and in Claude you can block the write tools entirely to keep the assistant read-only. An AI that answers customers on its own is the most common complaint about review software, so the default here is "the model drafts, a person approves".&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;2. Writes are idempotent where it matters.&lt;/strong&gt; A second &lt;code&gt;send_review_request&lt;/code&gt; for the same email at the same location returns &lt;code&gt;status: "already_requested"&lt;/code&gt;. No second email goes out and no credit is spent. Models retry; customers shouldn't get two emails because of it.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;3. Surface the platform's verdict, not just "success".&lt;/strong&gt; Google now moderates owner replies before showing them. So &lt;code&gt;publish_reply&lt;/code&gt; returns &lt;code&gt;google_state&lt;/code&gt;: &lt;code&gt;pending&lt;/code&gt; while Google checks, or &lt;code&gt;rejected_by_google&lt;/code&gt; with the policy reason. &lt;code&gt;list_reviews&lt;/code&gt; carries the same state, and a rejected reply counts as unanswered. Without that, an agent would happily report "replied!" to a reply nobody can see.&lt;/p&gt;

&lt;h2&gt;
  
  
  Prompts that work well
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;"Summarize what customers complained about at the Austin location in the last 30 days."&lt;/li&gt;
&lt;li&gt;"Draft replies to every unanswered 1–3 star review this week. Don't publish anything."&lt;/li&gt;
&lt;li&gt;"Compare average rating and unanswered count across all locations, worst first."&lt;/li&gt;
&lt;li&gt;"Send a review request to &lt;a href="mailto:jane@example.com"&gt;jane@example.com&lt;/a&gt; for the Round Rock location." (Claude asks first.)&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Costs and limits
&lt;/h2&gt;

&lt;p&gt;Each tool call costs one API credit (&lt;code&gt;get_usage&lt;/code&gt; is free), and every paid ReputeMap plan, including the 14-day trial, includes 1,000 credits a month, shared with the REST API.&lt;/p&gt;

&lt;h2&gt;
  
  
  Links
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Setup reference with Claude Desktop, Cursor and troubleshooting: &lt;a href="https://reputemap.com/docs/mcp" rel="noopener noreferrer"&gt;reputemap.com/docs/mcp&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;Copy-paste configs and scripts: &lt;a href="https://github.com/reputemap-com/examples" rel="noopener noreferrer"&gt;github.com/reputemap-com/examples&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;Agent skills for review triage and replies: &lt;code&gt;npx skills add reputemap-com/review-skills&lt;/code&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If you build something on top of it, I'd love to hear what you ask your reviews.&lt;/p&gt;

</description>
      <category>mcp</category>
      <category>ai</category>
      <category>claude</category>
      <category>api</category>
    </item>
  </channel>
</rss>
