Kaltura MCP Server

# 🎬 Kaltura MCP Server: Unleashing AI-Powered Video Intelligence Transform your Kaltura media library into an intelligent, searchable, and actionable knowledge base. This tutorial explores creative workflows that turn tedious manual video management into effortless AI-powered automation. ## 🚀 Quick Start: Your First AI Video Query After [setting up your MCP server](README.md), try this magic: **You**: *"Find my most recent training videos and summarize their content"* **Claude** will: 1. Search your Kaltura library for recent videos with "training" keywords 2. Retrieve captions/transcripts automatically 3. Analyze and summarize the content across multiple videos 4. Present insights you never knew existed in your library **Previously**: Hours of manual browsing, downloading, and watching videos **Now**: Seconds of intelligent discovery and analysis --- ## 🎯 Game-Changing Use Cases ### 📚 **Content Discovery & Intelligence** #### **"Find all videos mentioning machine learning from the last 6 months and create a learning roadmap"** ``` Claude automatically: - Searches across titles, descriptions, AND captions - Filters by date range - Extracts key concepts from transcripts - Organizes findings into a structured learning path - Suggests prerequisite relationships between videos ``` **Real Impact**: Transform scattered video content into curated learning experiences. #### **"Analyze user engagement patterns for our product demo videos"** ``` Claude intelligently: - Retrieves analytics for all demo videos - Identifies drop-off points and peak engagement - Correlates content topics with viewer retention - Suggests optimization strategies - Creates visual summaries of performance trends ``` **Real Impact**: Data-driven video strategy instead of guesswork. --- ### 🎥 **Content Production Workflows** #### **"Help me create a highlight reel from our conference presentations"** ``` Agentic workflow: 1. "Find all presentation videos from TechConf 2024" 2. "Extract key moments from transcripts where audience engagement was highest" 3. "Get thumbnails at those specific timestamps" 4. "Provide download URLs for the best segments" 5. "Create a production brief with timestamps and descriptions" ``` **Previously**: Days of manual review and note-taking **Now**: Minutes of AI-powered content curation #### **"Audit our video library for compliance and categorization"** ``` Intelligent automation: - Scans all videos for specific terminology or content types - Identifies uncategorized or miscategorized content - Flags videos missing captions for accessibility - Suggests metadata improvements - Creates compliance reports with actionable recommendations ``` **Real Impact**: Automated content governance at scale. --- ### 📊 **Analytics & Insights with Visual Dashboards** #### **"Create a visual performance dashboard for our marketing videos with daily trends"** ``` Claude orchestrates: 1. Retrieves graph data for all marketing category videos 2. Generates time-series charts showing: - Daily play counts and unique viewers - Average watch time trends - Engagement rate fluctuations 3. Creates visualization-ready data for Chart.js or Plotly 4. Identifies peak performance periods and anomalies 5. Suggests optimal posting times based on historical patterns ``` **Magic Moment**: Transform raw analytics into interactive visual dashboards with natural language. #### **"Show me viewer engagement trends for our product demos over the last quarter"** ``` Visualization workflow: - Fetches daily graph data for Q4 product demos - Returns formatted time-series data: { "graphs": [ {"metric": "count_plays", "data": [{"date": "2024-01-01", "value": 150}, ...]}, {"metric": "avg_completion_rate", "data": [{"date": "2024-01-01", "value": 0.75}, ...]} ] } - Highlights engagement patterns and seasonal trends - Provides data ready for immediate charting ``` **Real Impact**: Interactive dashboards that update with real-time data. #### **"Find our most engaging educational content and visualize performance patterns"** ``` Deep analysis with visualization: - Identifies highest-performing educational videos - Generates engagement timeline graphs showing: * Where viewers drop off vs. stay engaged * Peak interest moments throughout videos * Completion rate curves - Correlates content segments with engagement spikes - Creates visual heatmaps of viewer attention - Provides data for creating engagement dashboards ``` **Real Impact**: Visual evidence-based content strategy with clear performance indicators. #### **"Compare weekly performance of our training videos vs. product demos"** ``` Comparative visualization: - Retrieves weekly graph data for both content types - Returns multi-series comparison data: * Training videos: plays, completion rates, watch time * Product demos: plays, completion rates, watch time - Formats data for side-by-side chart comparison - Identifies which content type drives better engagement - Suggests resource allocation based on visual trends ``` **Visualization Output**: Ready-to-chart data comparing content performance over time. --- ### 🔍 **Advanced Search & Discovery** #### **"Find videos where [specific speaker] discusses [topic] and extract action items"** ``` Multi-modal intelligence: - Searches across metadata, captions, and descriptions - Identifies specific speaker appearances - Extracts topic-relevant segments - Summarizes key points and action items - Creates meeting notes and follow-up tasks ``` **Previously**: Impossible without manual review **Now**: Instant knowledge extraction from your video archives #### **"Create a competitive analysis from our recorded webinars"** ``` Strategic intelligence: 1. Identifies webinars mentioning competitors 2. Extracts competitive insights from transcripts 3. Analyzes trends in competitive positioning 4. Summarizes market insights across multiple sessions 5. Creates strategic briefing documents ``` **Real Impact**: Turn recorded content into competitive intelligence. --- ### 🎓 **Learning & Training Optimization** #### **"Optimize our employee onboarding videos based on completion rates"** ``` Data-driven optimization: - Analyzes completion rates across onboarding series - Identifies content segments with high drop-off - Correlates video length with engagement - Suggests content restructuring strategies - Recommends personalization approaches ``` **Magic**: Transform training effectiveness through video intelligence. #### **"Create personalized learning paths from our video library"** ``` Intelligent curation: 1. Analyzes user's role and experience level 2. Maps relevant videos from library 3. Sequences content for optimal learning progression 4. Identifies knowledge gaps and prerequisite content 5. Creates custom curricula with progress tracking suggestions ``` **Real Impact**: Personalized education at scale without manual curation. --- ## 🛠️ **Power User Workflows** ### **Multi-Video Analysis Pipeline** **Scenario**: *"Analyze all customer testimonial videos and create a comprehensive impact report"* ```markdown Step 1: Discovery "Search for all videos in the 'testimonials' category from this year" Step 2: Content Extraction "Get transcripts for all these videos and extract key themes" Step 3: Analytics Integration "Retrieve viewing analytics and engagement metrics for these videos" Step 4: Insight Generation "Identify most compelling customer benefits and success stories" Step 5: Report Creation "Create an executive summary with supporting data and recommendations" ``` **Result**: Comprehensive business intelligence from video content in minutes. ### **Content Lifecycle Management with Visual Analytics** **Scenario**: *"Audit and optimize our entire video content strategy with visual performance tracking"* ```markdown Phase 1: Content Inventory with Trends "List all videos and generate 30-day performance trend graphs for each category" Phase 2: Visual Performance Analysis "Create time-series charts showing: - Category performance over time - Viewer engagement patterns by content type - Upload frequency vs. engagement correlation" Phase 3: Content Gap Visualization "Generate graphs showing: - Topics with high search but low content - Engagement drops indicating missing follow-up content - Viewer journey visualization across content series" Phase 4: Optimization Dashboard "Create interactive dashboard data showing: - Content performance heatmaps - Optimal video length by category (graph data) - Best performing times and days (visual trends)" Phase 5: Predictive Planning "Generate forecast graphs based on historical trends" ``` **Result**: Visual content strategy with interactive dashboards and trend analysis. ### **Automated Content Workflows with Visual Reporting** **Scenario**: *"Set up intelligent content monitoring with visual weekly reports"* ```markdown Weekly Visual Intelligence Briefing: 1. "Find all new videos uploaded this week" 2. "Generate comparison graphs: - This week vs. last week performance - 4-week rolling average trends - Day-by-day engagement patterns" 3. "Create trend visualizations: - Rising topics (upward trending graphs) - Declining topics (downward trends) - Seasonal patterns in engagement" 4. "Visual alerts dashboard: - Red flag videos (steep engagement drops) - Green flag videos (exceptional performance) - Videos trending toward low engagement" 5. "Generate visual weekly report with: - Performance charts by category - Engagement timeline graphs - Predictive trend lines for next week" ``` **Result**: Visual content monitoring with trend-based predictive insights. --- ## 🚀 **NEW: Purpose-Based Analytics Functions (V2)** ### **Find the Right Analytics Tool Without Technical Knowledge** The new Analytics V2 provides intuitive, purpose-driven functions that make it easy to get exactly the analytics you need: #### **"I need to create a dashboard with charts"** ``` Use: get_analytics_timeseries() - Returns chart-ready data with consistent time intervals - Multiple metrics in one call - Direct integration with Chart.js, Plotly, D3.js ``` #### **"Show me where viewers stop watching"** ``` Use: get_video_retention() - 101-point retention curve for any video - Automatic drop-off and replay detection - Compare different viewer segments ``` #### **"Monitor our live event in real-time"** ``` Use: get_realtime_metrics() - Updates every ~30 seconds - Current viewers, geographic distribution, quality - Includes trends and peak detection ``` #### **"Analyze streaming quality issues"** ``` Use: get_quality_metrics() - Buffering rates, bitrate analysis, error tracking - Quality score and recommendations - Device and geographic breakdowns ``` #### **"Understand our global audience"** ``` Use: get_geographic_breakdown() - Country, state, or city-level analysis - Automatic percentage calculations - Top locations and growth insights ``` #### **"What analytics tools are available?"** ``` Use: list_analytics_capabilities() - Discover all functions and their purposes - See examples and use cases - Find the right tool for your needs ``` **Example Workflow**: Building a Complete Video Analytics Dashboard ``` 1. "List analytics capabilities" → Discover available functions 2. "Get performance rankings for last month" → get_analytics() 3. "Create daily play count chart" → get_analytics_timeseries() 4. "Analyze retention for top video" → get_video_retention() 5. "Check current live viewers" → get_realtime_metrics() 6. "Show geographic distribution" → get_geographic_breakdown() ``` **Migration Tip**: All old functions still work! New functions provide clearer naming and enhanced features. See the [Analytics V2 Guide](ANALYTICS_V2_GUIDE.md) for details. --- ## 📈 **Visual Analytics & Graph Data Capabilities** ### **Transform Analytics into Visual Stories** The Kaltura MCP now provides powerful graph data capabilities that transform raw analytics into visualization-ready time-series data. Here's how to leverage this for impactful insights: ### **Video Timeline Analytics - Granular Retention Analysis** #### **"Show me exactly where viewers drop off in my training video"** ``` Video Timeline Analytics (PERCENTILES): 1. Request: "Analyze retention for video 1_abc123" 2. Returns 101 data points (0-100% of video duration): - Exact viewer count at each percentile - Unique vs. repeat viewers - Replay hotspots identification 3. Insights generated: - Major drop-off points (e.g., "25% mark: 15% viewer loss") - Replay segments (e.g., "45% mark: 35% replay rate") - Average retention and completion rates ``` **Example prompt**: *"Get the retention curve for our product demo and identify where we lose viewers"* #### **"Compare how anonymous vs logged-in users watch our content"** ``` Cohort Comparison Workflow: - "Show retention curves comparing anonymous and registered viewers for video 1_xyz" - Returns: * Anonymous viewer retention curve * Logged-in user retention curve * Behavioral differences highlighted * Engagement score comparison ``` **Insight**: Discover if your content resonates differently with authenticated users. #### **"Find which parts of my tutorial get replayed the most"** ``` Replay Detection Analysis: - "Identify replay hotspots in video 1_tutorial" - Returns percentile-by-percentile replay data: * Segments with >20% replay rate flagged * Exact replay counts per segment * Suggestions for content optimization ``` **Use case**: Extract high-value segments for promotional clips or improve confusing sections. #### **"Create an engagement dashboard for our webinar series"** ``` Graph data workflow: 1. Request: "Get graph data for all webinars from the last 3 months" 2. Receive time-series data with multiple metrics: - Daily play counts - Average watch time trends - Completion rate patterns - Unique viewer growth 3. Data formatted for immediate visualization in: - Chart.js (web dashboards) - Plotly (interactive graphs) - D3.js (custom visualizations) - Excel/Google Sheets (business reports) ``` **Example prompt**: *"Generate daily engagement graphs for our 'Tech Talks' series showing plays, completion rates, and average watch time"* #### **"Visualize content performance patterns to optimize upload schedule"** ``` Pattern recognition through graphs: - "Show me hourly viewing patterns for weekdays vs weekends" - Returns graph data revealing: * Peak viewing hours by day * Engagement differences between content types * Optimal upload windows - Creates data for heat map visualizations - Identifies untapped time slots ``` **Visual insight**: See exactly when your audience is most engaged, not just aggregate numbers. #### **"Track campaign effectiveness with visual trend analysis"** ``` Campaign tracking workflow: 1. "Get daily graph data for videos tagged 'summer-campaign'" 2. Receive multi-metric time series: { "graphs": [ {"metric": "count_plays", "data": [...]}, {"metric": "unique_viewers", "data": [...]}, {"metric": "social_shares", "data": [...]} ], "summary": {"total_reach": 50000, "engagement_rate": 0.45} } 3. Track campaign momentum visually 4. Identify viral moments and plateau periods 5. Adjust strategy based on real-time trends ``` **Power move**: Real-time campaign optimization through visual trend monitoring. ### **Graph Data Prompt Examples** #### **Performance Tracking** - *"Graph our video performance metrics for the last 30 days with daily intervals"* - *"Show me weekly trends for our educational content vs entertainment content"* - *"Visualize the correlation between video length and completion rates"* #### **Video Timeline & Retention Analysis** - *"Show me the retention curve for video 1_abc123"* - *"Analyze where viewers drop off in our onboarding video"* - *"Find the exact points where people replay content in video 1_xyz"* - *"Compare retention patterns between mobile and desktop viewers"* - *"Identify which segments of our tutorial have the highest engagement"* #### **Cohort Comparison** - *"Compare retention curves: premium users vs free users for video 1_demo"* - *"Show how first-time viewers vs returning viewers watch our content"* - *"Analyze viewing patterns: employees vs external users"* - *"Compare anonymous viewer retention to logged-in users"* #### **Content Optimization** - *"Find the optimal video length based on retention data"* - *"Identify segments to cut based on drop-off analysis"* - *"Show me which intro length has the best retention"* - *"Analyze if our CTAs are causing viewer drop-off"* #### **Predictive Analytics** - *"Based on historical graphs, predict next month's viewing patterns"* - *"Show me seasonal trends in our video consumption over 2 years"* - *"Identify growth trajectories for different content types"* #### **Real-time Monitoring** - *"Get live viewer count graphs for our ongoing webinar"* - *"Show me real-time engagement metrics for today's product launch video"* - *"Track concurrent viewers throughout our live event"* ### **Creating Visual Reports** ``` Executive Dashboard Prompt: "Create a comprehensive visual analytics report for Q1: 1. Monthly trend graphs for all key metrics 2. Category performance comparison charts 3. Geographic distribution over time 4. Content ROI visualization 5. Predictive trends for Q2" Returns: Complete dataset formatted for executive dashboards ``` --- ## 🎨 **Creative Applications** ### **Content Repurposing Engine** **"Transform our webinar library into a content marketing machine"** ``` Intelligent repurposing: - Extracts key quotes and insights from webinar transcripts - Identifies quotable moments with timestamps - Suggests blog post topics from content themes - Creates social media content from engaging segments - Generates speaker spotlight materials ``` ### **Interactive Video Exploration** **"Create an interactive knowledge base from our training videos"** ``` Knowledge graph creation: - Maps relationships between video topics - Creates cross-references and related content suggestions - Builds searchable knowledge base from transcripts - Generates topic-based learning modules - Suggests content connections and pathways ``` ### **Predictive Content Planning with Visual Forecasting** **"Predict what video content will perform best next quarter with visual trend analysis"** ``` Visual trend analysis and prediction: - Generates historical performance graphs by content type - Creates seasonal trend visualizations: * Year-over-year comparison charts * Monthly pattern recognition graphs * Holiday impact on engagement metrics - Produces predictive charts showing: * Expected performance curves for Q2 * Optimal upload timing graphs * Content mix recommendations with projected ROI - Delivers forecast data formatted for dashboards: { "historical_trends": {"Q1_2023": [...], "Q1_2024": [...]}, "predictions": {"Q2_2024": [{"date": "2024-04-01", "predicted_plays": 250}, ...]}, "confidence_bands": {"upper": [...], "lower": [...]} } ``` **Visual Output**: Interactive forecast dashboards with confidence intervals. --- ## 🚀 **Getting Started: Your First Agentic Workflow** ### **Beginner: Content Discovery** 1. **Ask**: *"What are my 10 most popular videos this month?"* 2. **Follow up**: *"Analyze their transcripts to find common success factors"* 3. **Deep dive**: *"Create a content strategy based on these insights"* ### **Intermediate: Performance Optimization** 1. **Query**: *"Find videos with low engagement in my educational category"* 2. **Analyze**: *"Compare their content and structure to high-performing videos"* 3. **Action**: *"Suggest specific improvements for each underperforming video"* ### **Advanced: Strategic Intelligence** 1. **Research**: *"Analyze all customer-facing videos for sentiment and feedback themes"* 2. **Correlate**: *"Cross-reference with support ticket trends and product feedback"* 3. **Strategize**: *"Develop video content strategy to address customer pain points"* --- ## 💡 **Pro Tips for Maximum Impact** ### **🎯 Precision Querying with Visualization** - **Time-series analysis**: *"Show me daily play count trends for videos uploaded in the last 30 days"* - **Performance visualization**: *"Graph the completion rates over time for our training videos"* - **Engagement patterns**: *"Visualize hourly viewing patterns for our live event recordings"* - **Comparative charts**: *"Create a graph comparing viewer retention across our top 5 videos"* - **Geographic trends**: *"Show me a time-series of views by country for our global campaign"* ### **🔗 Workflow Chaining** - Start broad, then narrow: *Discovery → Analysis → Action* - Build on results: Use previous findings to guide next queries - Create feedback loops: *Performance → Insights → Strategy → Content → Performance* ### **📊 Visual Data-Driven Decisions** - **Request graph data**: *"Give me graph data for daily engagement over the last month"* - **Visualize trends**: *"Show me viewer count trends with weekly intervals"* - **Multi-metric charts**: *"Graph plays, completion rate, and average watch time together"* - **Pattern recognition**: *"Visualize engagement patterns to identify peak viewing times"* - **Dashboard creation**: *"Generate dashboard-ready data for our monthly review"* ### **📈 Example Graph Data Prompts** - **Performance Dashboard**: *"Create graph data for a performance dashboard showing daily metrics for our educational content"* - **Trend Analysis**: *"Generate time-series data showing monthly growth in viewer engagement"* - **A/B Testing**: *"Compare graph data for videos with different thumbnail styles"* - **ROI Visualization**: *"Show me the correlation between video length and completion rates as a graph"* - **Real-time Monitoring**: *"Get real-time viewer data for our live stream in graph format"* ### **🎬 Content Intelligence** - Leverage transcript analysis for deeper insights - Use thumbnail analysis for visual content strategies - Combine captions, metadata, and analytics for comprehensive understanding --- ## 🌟 **The Transformation Promise** **Before Kaltura MCP**: - Manual video browsing and searching - Disconnected analytics and content insights - Time-intensive content audits and analysis - Reactive content strategy - Siloed video knowledge **After Kaltura MCP**: - ✨ **Instant intelligent video discovery** - 🔍 **AI-powered content analysis and insights** - ⚡ **Automated video intelligence workflows** - 📊 **Visual analytics with interactive dashboards** - 📈 **Time-series data for trend visualization** - 🎯 **Graph-based performance tracking** - 🧠 **Unified video knowledge base** --- ## 🎪 **Real-World Success Stories** ### **Marketing Team**: *"We reduced content planning time by 80%"* *"Instead of manually reviewing hundreds of videos to plan our content calendar, Claude analyzes our entire library and suggests optimal content strategies based on actual performance data."* ### **Training Department**: *"Our course completion rates improved 40%"* *"By analyzing video engagement patterns, we identified exactly where learners were dropping off and optimized our content structure. What used to take weeks of analysis now happens in minutes."* ### **Content Creator**: *"I discovered goldmines in my own content"* *"Claude found connections and insights in my video library that I never noticed. It's like having a research assistant who never sleeps and remembers everything."* --- ## 🚀 **Ready to Transform Your Video Intelligence?** Your Kaltura library contains untapped intelligence waiting to be discovered. With the Kaltura MCP Server, every video becomes part of an intelligent, searchable, and actionable knowledge system. **Start your journey**: 1. Set up your MCP server using the [setup guide](README.md) 2. Try your first intelligent query 3. Experience the magic of AI-powered video intelligence 4. Share your discoveries and build upon the possibilities **The future of video management is intelligent, automated, and incredibly powerful. Welcome to the revolution.** --- *Ready to unlock the hidden intelligence in your video library? Your journey into AI-powered video management starts with a single question to Claude.*