Focusnote

Inspiration

The idea for Focusnote came from the growing challenge of remote school and work. Whether in online classes or professional meetings, it’s hard to stay engaged while trying to take detailed notes. Many students and workers struggle to listen attentively while also documenting key points, especially with constant digital distractions. We wanted to design something that lets users focus on the conversation instead of their keyboards.

What it does

Focusnote is a desktop meeting assistant powered by AI that transcribes conversations in real-time, summarizes discussions, extracts action items, and organizes everything automatically. It detects when you’re in a call or online lecture, activates “Do Not Disturb” mode, records and transcribes speech, and saves all content in a searchable meeting and class notes hub. It integrates seamlessly with Zoom, Microsoft Teams, Google Meet, and Discord, making it perfect for both school and workplace settings.

How we built it

We built Focusnote using Python as the main programming language. Python’s strong ecosystem for AI and natural language processing allowed us to use speech-to-text APIs for real-time transcription and AI summarization models to extract key points and action items. The backend handles call detection, and transcription, while the frontend displays summaries in an organized dashboard. Focusnote was designed to support both academic environments (lectures, group projects) and professional use (team meetings, client discussions) efficiently.

Challenges we ran into

getting a frequency where the audio sounded normal. being able to transcribe data live while streaming audio. finding a good length to segment the audio data to send to the transcription model. able to get system audio period.

Accomplishments that we're proud of

Fully automated workflow from class or meeting detection to note generation. Seamless integration with educational and business communication tools. Accurate AI summaries that capture lectures and meetings clearly. A user experience that allows students and professionals to stay fully focused. Building something truly useful for learning and productivity.

What we learned

We learned how to combine Python, AI, speech recognition, and design thinking to make a tool that enhances both education and professional work. We also realized how important it is to protect users’ privacy and create summaries that people can trust whether they’re revising for exams or following up on business decision

What's next for Focusnote

Adding multi-language support for international students and global teams. Creating mobile and browser extensions for accessibility anywhere. Enhancing custom AI models to adapt to academic or corporate meeting styles. Building shared class and team workspaces for collaboration. Expanding integrations with learning and productivity tools like Google Classroom, Notion, Slack, and Microsoft 365.

Built With

• ffmpeg
• gemini
• python
• pywhispercpp
• websockets