Fathom

• 

  Fathom Landing Page

Inspiration As students ourselves, we often struggled to capture every essential detail during fast-paced lectures. Fathom was born out of the need for a smarter, more efficient way to retain lecture content without the added stress of note-taking. We envisioned a tool that could bridge this gap, allowing students to focus on understanding rather than transcription and make learning more accessible by simplifying the review process.

What We Learned Building Fathom taught us valuable skills in real-time data processing and backend development, particularly in integrating transcription models with web applications. We also deepened our understanding of natural language processing (NLP) for text summarization, improving our ability to extract key points from large text bodies. Collaboration played a huge role in our learning experience, as we tackled problem-solving, teamwork, and agile development practices.

How We Built Fathom The app is built using a robust backend powered by Python, where we integrate the Fast Whisper model for real-time transcription. Our frontend is designed with Next.js, offering a seamless user experience where students can upload their lecture recordings, receive a full transcript, and review key points generated by NLP-based summarization. Together, these components create a streamlined and intuitive user interface that emphasizes ease of use.

Challenges Throughout development, we faced multiple challenges. Implementing real-time transcription required precise tuning to balance speed and accuracy, especially when processing lengthy audio files. We also encountered integration issues with some dependencies and had to troubleshoot unexpected bugs, such as those in handling simultaneous audio uploads. Despite these challenges, each hurdle strengthened our skills and commitment to making Fathom as efficient and effective as possible.

What's Next Looking forward, we aim to enhance Fathom by adding multi-language support and more advanced summarization features that allow users to customize the level of detail in the generated key points. We’re also exploring integrations with popular learning management systems (LMS) so students can directly import lecture recordings from platforms like Zoom or Blackboard. Finally, we hope to introduce collaborative features where users can share and discuss important lecture points within study groups, transforming Fathom into an interactive tool for collective learning.

Built With

• amazon-web-services
• api
• fast-whisper
• javascript
• mongodb
• natural-language-processing
• nextjs
• python