<?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: Youla Lee</title>
    <description>The latest articles on DEV Community by Youla Lee (@battlelolo).</description>
    <link>https://dev.to/battlelolo</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.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3455352%2F2330c31c-03ff-4463-b90a-107bebd36394.png</url>
      <title>DEV Community: Youla Lee</title>
      <link>https://dev.to/battlelolo</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/battlelolo"/>
    <language>en</language>
    <item>
      <title>Building a TTS Tool for My Friend in One Hour</title>
      <dc:creator>Youla Lee</dc:creator>
      <pubDate>Sun, 24 Aug 2025 03:34:54 +0000</pubDate>
      <link>https://dev.to/battlelolo/building-a-tts-tool-for-my-friend-in-one-hour-k6i</link>
      <guid>https://dev.to/battlelolo/building-a-tts-tool-for-my-friend-in-one-hour-k6i</guid>
      <description>&lt;h1&gt;
  
  
  Building a TTS Tool for My Friend in One Hour
&lt;/h1&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fk3ud1x6nymg5ji91yw2q.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fk3ud1x6nymg5ji91yw2q.png" alt=" " width="756" height="644"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The Ask
&lt;/h2&gt;

&lt;p&gt;My friend asked if there was a service that could read academic papers aloud - not like NotebookLM which creates podcast-style summaries, but something that would actually read the original text. She wanted to listen to papers like audiobooks when her eyes got tired.&lt;/p&gt;

&lt;p&gt;I didn't know of such a service, but since I'm familiar with Microsoft Azure Language Services, I offered to help: "Send me the paper and I'll make mp3 for you."&lt;/p&gt;

&lt;h2&gt;
  
  
  The Reality Check
&lt;/h2&gt;

&lt;p&gt;I thought this would be simple:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Extract text from PDF using Claude/ChatGPT/Grok
&lt;/li&gt;
&lt;li&gt;Run it through Azure TTS (Text to Speech)&lt;/li&gt;
&lt;li&gt;Done!&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Wrong. Academic PDFs are messy. Extract text and you get dozens of co-author names, chart numbers, table data, footnotes - everything my friend didn't want to hear.&lt;/p&gt;

&lt;p&gt;I tried asking different AIs to extract only title, abstract, and main content:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Claude Sonnet: Failed&lt;/li&gt;
&lt;li&gt;Grok: Failed with errors
&lt;/li&gt;
&lt;li&gt;Claude Opus: Success&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Turns out extracting clean content from academic PDFs is harder than expected.&lt;/p&gt;

&lt;h2&gt;
  
  
  From Manual to Automated
&lt;/h2&gt;

&lt;p&gt;I generated the MP3 using Azure TTS and made a tutorial video. But a week later, my friend was still hunting for paid PDF-to-speech services.&lt;/p&gt;

&lt;p&gt;That's when I realized: I might not know the perfect PDF-to-text AI, but I can definitely build a basic TTS web app. &lt;/p&gt;

&lt;p&gt;So I coded one with Claude Sonnet, built it in Next.js, deployed on Vercel, and shared it.&lt;/p&gt;

&lt;p&gt;This is why I love being able to code!!!&lt;/p&gt;




&lt;p&gt;Visit my github blog to read more&lt;br&gt;
&lt;a href="https://battlelolo.github.io/" rel="noopener noreferrer"&gt;&lt;/a&gt;&lt;a href="https://battlelolo.github.io/" rel="noopener noreferrer"&gt;https://battlelolo.github.io/&lt;/a&gt;&lt;/p&gt;

</description>
      <category>tts</category>
      <category>azure</category>
      <category>ai</category>
      <category>webdev</category>
    </item>
  </channel>
</rss>
