The example projects on this page utilize the TwelveLabs Video Understanding Platform to unlock hidden insights and enable precise content retrieval.

TrailSense

Summary: TrailSense is a natural language video search engine tailored specifically for mountain biking trails. It assists riders in discovering trail characteristics through “vibe-based” queries, such as “fast desert trail with berms.”

Description: The platform utilizes the TwelveLabs API for core video indexing along with Marengo model processing and semantic search capabilities. It also incorporates Gemini for conversational AI responses and automated difficulty ratings. This specialized application addresses real challenges in outdoor recreation discovery by allowing descriptive searches that traditional trail maps and ratings cannot provide.

GitHub repo: qZheng/trailsense, UmerQureshi21/TrailSenseV1

Devpost: TrailSense

LifeOS

Summary: LifeOS creates a continuous “digital memory” system that automatically captures, processes, and organizes life experiences through intelligent video analysis.

Description: The platform demonstrates exceptional TwelveLabs integration using both Marengo 2.7 for video embeddings and Pegasus 1.2 for summarization, implementing real-time processing pipelines with semantic search capabilities across personal video history. This innovative system addresses AI’s fundamental limitation of lacking personal context by building automated workflow triggers and memory assistance based on continuous video capture

GitHub repo: owenguoo/LifeOS

Devpost: LifeOS