The latest articles on Forem by prasanna-lakshmi18 (@prasannalakshmi18). https://forem.com/prasannalakshmi18 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%2F1612684%2Fce8ad539-f460-4a5b-95e7-ae1c8928ce07.png https://forem.com/prasannalakshmi18 en prasanna-lakshmi18 Mon, 01 Sep 2025 06:46:09 +0000 https://forem.com/prasannalakshmi18/ai-agent-for-youtube-comment-analysis-video-suggestions-2dpn https://forem.com/prasannalakshmi18/ai-agent-for-youtube-comment-analysis-video-suggestions-2dpn <p><em>This is a submission for the <a href="https://dev.to/challenges/brightdata-n8n-2025-08-13">AI Agents Challenge powered by n8n and Bright Data</a></em></p> <h2> What I Built </h2> <p>I built an AI Agent using <strong>n8n</strong> and <strong>Bright Data</strong> that analyzes YouTube comments and generates personalized video suggestions and insights.<br><br> The goal is to help content creators or marketers understand audience sentiment, find new content opportunities, and improve engagement strategies.</p> <h2> Demo </h2> <ul> <li><a href="https://prasannamallikarjun.app.n8n.cloud/workflow/AYDdRjXUsVq5E0Wh" rel="noopener noreferrer">Live Demo</a></li> </ul> <h2> n8n Workflow </h2> <p>I’ve shared the full workflow JSON on GitHub:<br><br> 👉 <a href="https://github.com/prasanna-lakshmi18/n8n-ai-agent-with-bright-data/blob/main/Youtube_comment_analysis.json" rel="noopener noreferrer">View Workflow on GitHub</a></p> <h2> Technical Implementation </h2> <p>Here’s how the workflow works step by step:</p> <ol> <li> <strong>Trigger</strong> → A Schedule Trigger node runs once every week. </li> <li> <strong>Bright Data Scraper (1st Node)</strong> → Initiates a scrape of YouTube video comments, generating a <code>snapshotId</code>. </li> <li> <strong>Bright Data Scraper (2nd Node)</strong> → Uses the <code>snapshotId</code> to download the comments data. </li> <li> <strong>JavaScript Code Node</strong> → Parses and cleans the JSON comment data. </li> <li> <strong>Gemini AI Node</strong> → Takes the extracted comments and generates: <ul> <li>Engagement strategies </li> <li>Content improvement tips </li> </ul> </li> <li> <strong>Email Node</strong> → Sends the processed results to the intended recipient automatically.</li> </ol> <h2> Bright Data Verified Node </h2> <p>I used Bright Data’s <strong>Verified Node</strong> to scrape YouTube comment sections efficiently and reliably. It provides structured JSON data (via <code>snapshotId</code>), which I could easily process inside n8n for further AI analysis.</p> <h2> Journey </h2> <p>This project taught me how to integrate:</p> <ul> <li> <strong>Bright Data scraping</strong> → for reliable comment extraction </li> <li> <strong>n8n workflow automation</strong> → to orchestrate the entire process </li> <li> <strong>Gemini AI</strong> → to provide actionable insights </li> <li> <strong>Email automation</strong> → for timely delivery of results </li> </ul> <p>Challenges I overcame:</p> <ul> <li>Parsing large YouTube comment datasets in n8n </li> <li>Handling async data retrieval with Bright Data’s snapshot mechanism </li> <li>Formatting the AI-generated output in a useful way for creators </li> </ul> <p>I learned how powerful combining <strong>data scraping + automation + AI</strong> can be in real-world scenarios.</p> <p><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%2Fperesrhbeevhrsat1fae.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%2Fperesrhbeevhrsat1fae.png" alt="n8n youtube comment analyzer and suggestions ai agent with bright data demonstration screeshot" width="800" height="271"></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%2F843mgbj5e3qtqvbzb3mt.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%2F843mgbj5e3qtqvbzb3mt.png" alt="Output" width="800" height="450"></a></p> devchallenge n8nbrightdatachallenge ai webdev prasanna-lakshmi18 Mon, 11 Aug 2025 07:05:28 +0000 https://forem.com/prasannalakshmi18/ai-powered-study-companion-for-gate-aspirants-with-redis-8-2606 https://forem.com/prasannalakshmi18/ai-powered-study-companion-for-gate-aspirants-with-redis-8-2606 <p><em>This is a submission for the <a href="https://dev.to/challenges/redis-2025-07-23">Redis AI Challenge</a>: Beyond the Cache</em>.</p> <h2> What I Built </h2> <p>I built an <strong>AI-powered study companion</strong> tailored for GATE aspirants. The app goes beyond being a simple chatbot—it provides:</p> <ul> <li> <strong>Real-time question answering</strong> from a curated syllabus-based knowledge base.</li> <li> <strong>Audio-based interaction</strong> for hands-free learning, using streaming voice input and output.</li> <li> <strong>Personalized recommendations</strong> for study topics based on previous performance.</li> <li> <strong>Semantic search</strong> over vast study materials.</li> <li> <strong>Daily quizzes</strong> with adaptive difficulty to target weak areas.</li> </ul> <p>The goal was to create an <strong>engaging, accessible, and always-learning AI tutor</strong> that adapts to each student’s needs.</p> <h2> Demo </h2> <p>🎥 <strong>Video demo:</strong> <a href="https://github.com/prasanna-lakshmi18/Redis-challenge" rel="noopener noreferrer">here</a><br> 🖥 <strong>Live app:</strong><a href="https://github.com/prasanna-lakshmi18/Redis-challenge" rel="noopener noreferrer">App here</a></p> <p>For more information visit my github repo:<a href="https://github.com/prasanna-lakshmi18/Redis-challenge" rel="noopener noreferrer">Repo</a></p> <h2> How I Used Redis 8 </h2> <p>Redis was the <strong>real-time backbone</strong> of this app, not just a cache:</p> <ol> <li> <p><strong>Vector Search with Redis Stack</strong> </p> <ul> <li>I stored embeddings of GATE syllabus topics, lecture notes, and past question papers using Redis Vector Search.</li> <li>This enabled fast semantic search for relevant concepts when a user asked a question.</li> </ul> </li> <li> <p><strong>Streams for Live Audio Interaction</strong> </p> <ul> <li>Redis Streams handled live audio data from the user and sent it to the AI model in real-time.</li> <li>This powered a streaming Q&amp;A mode without lag.</li> </ul> </li> <li> <p><strong>Pub/Sub for Real-time Quiz Updates</strong> </p> <ul> <li>When a student takes a quiz, results are published instantly, and other connected sessions (like progress dashboards) update live.</li> </ul> </li> <li> <p><strong>Hash &amp; JSON Structures for User Profiles</strong> </p> <ul> <li>RedisJSON stored student profiles, past performance metrics, and preferences.</li> <li>This allowed the AI to dynamically tailor study recommendations.</li> </ul> </li> <li> <p><strong>Time-series Data for Progress Tracking</strong> </p> <ul> <li>Using RedisTimeSeries, I logged study sessions, quiz scores, and topic completion rates.</li> <li>This powered analytics to help students visualize their improvement over time.</li> </ul> </li> </ol> <p>By combining <strong>Vector Search</strong>, <strong>Streams</strong>, <strong>Pub/Sub</strong>, <strong>JSON</strong>, and <strong>TimeSeries</strong>, I turned Redis 8 into the <strong>primary real-time data layer</strong> for the AI tutor—far beyond simple caching.</p> redischallenge devchallenge database ai