The latest articles on DEV Community by Varun Dasharadhi (@varundasharadhi). https://dev.to/varundasharadhi 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%2F3910370%2Fd21a7290-afa0-4dc9-87d1-96720a656d3c.png https://dev.to/varundasharadhi en Varun Dasharadhi Sun, 03 May 2026 13:33:04 +0000 https://dev.to/varundasharadhi/how-i-built-an-emotion-reading-ai-in-24-hours-using-claude-hume-evi-3ggf https://dev.to/varundasharadhi/how-i-built-an-emotion-reading-ai-in-24-hours-using-claude-hume-evi-3ggf <h2> The Idea </h2> <p>What if AI could actually feel what you're feeling — <br> not just read your words?</p> <p>That was the spark behind EmpathIQ. Built solo in 24 <br> hours for the Replit 10 Buildathon.</p> <h2> What It Does </h2> <p>EmpathIQ combines:</p> <ul> <li>👁️ Facial emotion detection via webcam</li> <li>🎙️ Vocal emotion analysis via Hume EVI</li> <li>🤖 Claude API responses calibrated to both signals</li> </ul> <p>The result? An AI that responds to how you actually <br> feel — not just what you type.</p> <h2> The Feature That Surprised Me Most </h2> <p>Smart Glasses mode 🥽</p> <p>Point the camera at someone ELSE. EmpathIQ reads <br> THEIR emotion and gives YOU real-time coaching <br> on what to say.</p> <p>Angry person in front of you?<br> → "Lower your voice and acknowledge their concern <br> without arguing"</p> <p>The future vision is Meta Ray-Ban integration — <br> real-time emotional coaching in every room you <br> walk into.</p> <h2> The Tech </h2> <ul> <li>React + Vite</li> <li>face-api.js — facial emotion detection</li> <li>Hume EVI — vocal emotion AI</li> <li>Claude API — emotionally calibrated responses</li> <li>Recharts — emotion timeline chart</li> <li>Tailwind CSS</li> </ul> <h2> The Hardest Part </h2> <p>Combining two real-time emotion signals (face + voice) <br> into one coherent reading without lag or conflicts. <br> The fusion panel took several iterations to get right.</p> <h2> What I'd Do Differently </h2> <p>Start with the voice mode earlier — EVI integration <br> took longer than expected and nearly didn't make <br> the 24hr deadline.</p> <h2> What's Next </h2> <ul> <li>Apple Watch pulse + biometric fusion</li> <li>Meta smart glasses integration</li> <li>Clinical/therapy version (HIPAA compliant)</li> <li>Mobile app</li> </ul> <h2> Try It </h2> <p>🔗 Live: <a href="https://empathiq-studio--varundasharadhi.replit.app" rel="noopener noreferrer">https://empathiq-studio--varundasharadhi.replit.app</a><br> 🎬 Demo: <a href="https://www.loom.com/share/ee3177d34b40404487115fca5f8366ed" rel="noopener noreferrer">https://www.loom.com/share/ee3177d34b40404487115fca5f8366ed</a><br> ⭐ GitHub: <a href="https://github.com/VarunDasharadhi/Empathiq-Studio" rel="noopener noreferrer">https://github.com/VarunDasharadhi/Empathiq-Studio</a></p> <p>Would love your feedback! 🙏<br> <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%2Fzjat9urae8uhap8my32y.png" class="article-body-image-wrapper"><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%2Fzjat9urae8uhap8my32y.png" alt=" " width="800" height="508"></a><br> <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%2Flmbgefy9va2msu896jtb.png" class="article-body-image-wrapper"><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%2Flmbgefy9va2msu896jtb.png" alt=" " width="800" height="512"></a></p> ai buildinpublic react showdev