<?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: Bounce Layer</title>
    <description>The latest articles on DEV Community by Bounce Layer (@bouncelayer).</description>
    <link>https://dev.to/bouncelayer</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%2F4251965%2F6072e04e-8894-422a-9e02-b34f77a02c15.png</url>
      <title>DEV Community: Bounce Layer</title>
      <link>https://dev.to/bouncelayer</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/bouncelayer"/>
    <language>en</language>
    <item>
      <title>Let your AI agent check an email address before it sends</title>
      <dc:creator>Bounce Layer</dc:creator>
      <pubDate>Wed, 18 Nov 2026 13:20:04 +0000</pubDate>
      <link>https://dev.to/bouncelayer/let-your-ai-agent-check-an-email-address-before-it-sends-2lg7</link>
      <guid>https://dev.to/bouncelayer/let-your-ai-agent-check-an-email-address-before-it-sends-2lg7</guid>
      <description>&lt;p&gt;We build BounceLayer, an email verification service, so this post uses our MCP server. The same idea works with any verification server, and we list the others at the end.&lt;/p&gt;

&lt;h2&gt;
  
  
  The problem
&lt;/h2&gt;

&lt;p&gt;Agents that send email will happily send to an address that doesn't exist. Each one bounces, and enough bounces push a domain's mail into spam. A person would check the address first. An agent needs a tool for that.&lt;/p&gt;

&lt;h2&gt;
  
  
  The setup
&lt;/h2&gt;

&lt;p&gt;An MCP server gives the agent a &lt;code&gt;verify_email&lt;/code&gt; tool. With Claude Code it is one command:&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 bouncelayer https://api.bouncelayer.com/mcp/ \
  --header &lt;span class="s2"&gt;"X-API-Key: bl_live_YOUR_KEY"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;For Cursor or any other MCP client, run the npm package:&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;"bouncelayer"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
      &lt;span class="nl"&gt;"command"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"npx"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
      &lt;span class="nl"&gt;"args"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"-y"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"@bouncelayer/mcp"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
      &lt;span class="nl"&gt;"env"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="nl"&gt;"BOUNCELAYER_API_KEY"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"bl_live_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;h2&gt;
  
  
  No key yet
&lt;/h2&gt;

&lt;p&gt;The agent can get its own. The server has a &lt;code&gt;create_agent_account&lt;/code&gt; tool that needs no key and returns one with 25 free credits. It also returns a claim link for the person the agent works for. Claiming the account gives 300 free checks a month.&lt;/p&gt;

&lt;h2&gt;
  
  
  What comes back
&lt;/h2&gt;

&lt;p&gt;Ask the agent to check an address and it gets one of six results:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Result&lt;/th&gt;
&lt;th&gt;Meaning&lt;/th&gt;
&lt;th&gt;What the agent should do&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;valid&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;The mailbox exists&lt;/td&gt;
&lt;td&gt;Send&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;invalid&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;It doesn't exist&lt;/td&gt;
&lt;td&gt;Don't send&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;catch_all&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;The domain accepts everything&lt;/td&gt;
&lt;td&gt;Send with care&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;disposable&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;A throwaway address&lt;/td&gt;
&lt;td&gt;Don't send&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;do_not_mail&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;A risky address, such as a shared inbox&lt;/td&gt;
&lt;td&gt;Leave it out&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;unknown&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;The mail server gave no answer&lt;/td&gt;
&lt;td&gt;Try again later (not charged)&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Each result also has an &lt;code&gt;explanation&lt;/code&gt; in plain English and a &lt;code&gt;did_you_mean&lt;/code&gt; suggestion when the domain looks mistyped, so the agent can tell the user why and offer the fix.&lt;/p&gt;

&lt;h2&gt;
  
  
  A prompt that works
&lt;/h2&gt;

&lt;blockquote&gt;
&lt;p&gt;Before you send any email, call verify_email on the recipient. Only send if the result is valid. If it is catch_all, ask me first. If did_you_mean is set, show me the suggestion.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;h2&gt;
  
  
  Lists
&lt;/h2&gt;

&lt;p&gt;For more than a few addresses, use &lt;code&gt;submit_bulk_verification&lt;/code&gt; and then &lt;code&gt;get_bulk_job&lt;/code&gt; to collect the results. Checking a list one address at a time works, but it is slower and uses more of the agent's context.&lt;/p&gt;

&lt;h2&gt;
  
  
  Other servers
&lt;/h2&gt;

&lt;p&gt;ZeroBounce, Hunter, Bouncer, Clearout, MyEmailVerifier, EmailListVerify and EmailVerify.io also publish MCP servers. If you already pay one of them, use theirs. We compared how each one connects here: &lt;a href="https://bouncelayer.com/compare/email-verification-mcp-servers" rel="noopener noreferrer"&gt;https://bouncelayer.com/compare/email-verification-mcp-servers&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Docs for ours: &lt;a href="https://bouncelayer.com/docs/mcp" rel="noopener noreferrer"&gt;https://bouncelayer.com/docs/mcp&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Source (MIT): &lt;a href="https://github.com/BounceLayer/bouncelayer-mcp" rel="noopener noreferrer"&gt;https://github.com/BounceLayer/bouncelayer-mcp&lt;/a&gt;&lt;/p&gt;

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