GooseMeet

Keeping your team on track and your meetings in sync—one honk at a time!

Comment

Inspiration 🪿

The dreaded morning scrum. Endless meetings, forgotten action items, and the constant repetition of questions that were answered yesterday or the week before. As co-op students ourselves, we understand the struggle all too well. That’s why we created a way to simplify meetings and keep everyone on track.

What it does 🦆

Introducing GooseMeet, a wholesome intelligent little goose 🪿 that accompanies your team through all your meetings. As the perfect meeting companion, Mr. Goose 🪿tracks every meeting, remembers past meetings, and has the intelligence to provide insights based on past discussions.

Unlike other meeting assistants which provide you with insights/data after the meeting has ended, Mr. Goose 🪿 will talk to you in real-time using state-of-the-art voice models, and help make your meetings more fruitful! He can answer any questions you may have, such as: “what was discussed in yesterday’s meeting,” “which intern accidentally pushed to main last Friday,” or even, "help us brainstorm marketing ideas!"

Helping your team stay organized, Mr. Goose ensures no detail slips through the cracks, and that your meetings stay efficient and productive.

How we built it 🔧

  • HTML/CSS + JS, React for frontend
  • Flask backend with Python
  • WebRTC and WebSockets for real-time communication
  • Goose model originally created by @Atlasmodels on Sketchfab, licensed via CC by 4.0
    • Edited model in Blender to add: custom wings, upscaled and modified topology, and animations
  • Groq for Speech-to-Text capabilities
  • Cohere for LLM capabilities
  • Pinecone for vector database (to store embeddings of the conversational data)
  • Langchain for RAG (retrieving and understanding context from past meetings/conversations)

Challenges we ran into 🏁

  • Setting up WebSockets and RTC
  • Building an entire meeting app and resolving RTC issues
  • Bringing Mr. Goose to life through 3-D modeling, animation and minimizing latency

Accomplishments that we're proud of 🌟

  • Setting up real-time communication (RTC) to livestream audio and video between users
  • Implementing OpenAI and Groq for text-to-speech with minimal latency for realistic conversations
  • Integrating all the parts together

What we learned 📖

  • Designing and animating Mr. Goose in 3D using Three.js and Blender
  • How to build a Zoom-style meeting application in the span of a weekend, and attach real-time AI/RAG and goose animation capabilities to it

What's next for GooseMeet 📈

  • Teach Mr. Goose how to type in the chat
  • Teach Mr. Goose how to read documents
  • A dashboard to review the brain and the brawn (the metrics and the data) behind Mr. Goose.
  • Allow Mr. Goose to travel to Google Meets, Zoom, and Microsoft Teams (as an extension)

Built With

Share this project:

Updates