HootHive

AI-powered learning assistant that transforms recorded lectures into structured, accessible, and collaborative study resources:

Comment 1

Inspiration

Ever felt buried under hours of lectures, struggling to find that one concept right before an exam? College students today face information overload, low retention, fragmented study tools, language barriers, and limited collaboration opportunities. We wanted to solve a simple but universal problem: how can students turn dense, unstructured lectures into clear, engaging, and collaborative study resources? That’s how HootHive was born — an AI-powered learning assistant that transforms recorded lectures into structured, accessible, and shareable knowledge.

What it does

HootHive helps students learn smarter, not harder. With just a video upload, students can:

  • Summarize lectures with timestamps – Quickly review the key points without scrubbing through hours of content.

  • Chat with lecture videos (Q&A) – Ask questions and get precise snippet retrieval from the video.

  • Flashcards & Quizzes – Automatically generate practice materials on the fly, downloadable as PDFs.

  • Community Collaboration – Share notes, snippets, and study PDFs through our integrated HootBot (Discord bot).

Unlike generic LLMs like ChatGPT, HootHive is purpose-built for video understanding, bridging the gap between platforms like YouTube, Quizlet, and note-sharing apps into a single ecosystem.

How we built it

  • Backend: Python FastAPI orchestrates the data pipeline.

  • AI Core: At the heart of HootHive is the TwelveLabs AI Platform (SDK v1.0.2+), which provides multimodal video understanding capabilities. We leverage their Marengo 2.7 engine for video embeddings and Pegasus 1.2 for summarization, enabling automatic chapter detection, highlight extraction, and natural language video search.

  • Video Processing: We implemented robust video handling using MoviePy for video manipulation and segment creation, combined with ffmpeg for HLS stream processing. This enables precise video snippet extraction based on AI-generated timestamps.

  • Frontend: A clean and interactive UI built with Streamlit.

  • Study Material Production: We generate comprehensive study guides as PDFs using Gemini for document creation

  • Collaboration:🤖 Discord Integration We built a dedicated FastAPI microservice (discord_bot_server.py) that handles Discord bot communication. In our case, HootBot, which enables seamless sharing of AI-generated study content

Our pipeline is designed to be domain-agnostic — working across STEM, humanities, and even interdisciplinary courses.

Challenges we ran into

  • Integrating multiple AI models into a unified, low-latency pipeline.

  • Handling long-form videos while maintaining accurate snippet retrieval.

  • Ensuring cross-platform usability (web + Discord bot) without fragmenting the student experience.

  • Ensuring cross-domain proficiency ability to retrieve + generate accurate info for all scientific domains.

Accomplishments that we're proud of

  • Built a functioning software that can be scaled to production level, and that turns any lecture video into summaries, quizzes, and flashcards in minutes.

  • Seamlessly connected the ecosystem with HootBot for collaborative study groups.

  • Demonstrated cross-domain success — tested on Chemistry, Math, Biology, and even Arts lectures.

What we learned

  • Building with video-based RAG is significantly harder than text-only retrieval, but the payoff in accuracy and contextual understanding is worth it.

  • Students value collaboration tools just as much as study tools — community features are key to adoption.

  • Designing for accessibility and inclusivity is critical for global student impact.

What's next for HootHive

  • Cross-video search: Query across multiple uploaded lectures to connect concepts.

  • Personalized study recommendations: Generate tailored learning paths from YouTube, MOOCs, or student-uploaded content.

  • Scalability: Deploying as a SaaS platform for universities and student organizations.

  • Monetization: Freemium model (core features free, advanced analytics & collaborative workspaces as premium).

Built With

Share this project:

Updates