The latest articles on DEV Community by AttentionIsImportant (@attentionisalluneed). https://dev.to/attentionisalluneed 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%2F4107069%2Fadd01e25-18a6-47f2-8dcb-36a87640e88b.png https://dev.to/attentionisalluneed en AttentionIsImportant Thu, 03 Sep 2026 05:05:50 +0000 https://dev.to/attentionisalluneed/turn-markdown-into-narrated-videos-with-a-new-ai-skill-1i31 https://dev.to/attentionisalluneed/turn-markdown-into-narrated-videos-with-a-new-ai-skill-1i31 <p>I’ve been trying out a new Skill that turns Markdown content into narrated videos — and it’s surprisingly practical. The workflow is simple.</p> <p>The biggest advantage is that it doesn’t just generate slides or audio separately — it can directly produce a <strong>complete video with synchronized audio</strong>.</p> <p>In other words, the final output is an actual MP4 video with the generated narration already included:</p> <p><strong>Markdown → Slides + Audio → MP4 Video</strong></p> <p>You can also modify and build on top of the Skill to fit your own workflow. It is released under the <strong>Apache License 2.0</strong>, so you can customize, extend, and integrate it into your own projects, as long as you comply with the license requirements, including preserving the required attribution and license notices.</p> <p>You write your content in Markdown, and the Skill handles the rest:</p> <ul> <li>Converts Markdown into presentation slides with Marp</li> <li>Generates narration with Edge-TTS (Text-to-Speech)</li> <li>Synchronizes the slides and narration</li> <li>Uses FFmpeg to produce the final MP4 video</li> <li>Keeps the whole workflow local and scriptable</li> </ul> <p>This is particularly useful for turning technical content into:</p> <ul> <li>Programming tutorials</li> <li>Course materials</li> <li>Documentation</li> <li>AI-generated educational content</li> <li>Internal technical training</li> <li>Short-form technical videos</li> </ul> <p>The interesting part is that the input remains just Markdown.</p> <p>That means an AI agent can generate the Markdown first, and the Skill can take it from there.</p> <p>For example:<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight markdown"><code><span class="gh"># What is a Transformer?</span> <span class="gu">## Self-Attention</span> Self-attention allows each token to consider other tokens in the sequence... <span class="gu">## Why does it matter?</span> It enables the model to capture relationships between distant tokens. </code></pre> </div> <p>→ slides<br> → narration<br> → synchronized video</p> <p>No video editor required.</p> <p>The Skill is available here:</p> <p><a href="https://github.com/70v-Yoyo/md2video-audio-skill" rel="noopener noreferrer">https://github.com/70v-Yoyo/md2video-audio-skill</a></p> <p>If you regularly turn Markdown, documentation, or technical notes into educational content, this is worth trying.</p> ai skill markdown agentskills