Horoscrypt

Inspiration

As the MS Teams doesn’t provide API for transcription on their free plan, we built it for ourselves. Our team enthusiasm.

What it does

Introducing a webapp, that gets transcripts from given url of recorded Teams video. With the help of natural language processing, form spoken topics into a summarized text. Based on these topics, the AI will create a sort of To-Do List consisting future mentions discussed in the meeting. It will also calculate participants stratification index, and pass this information to a pre-Filled email which is available for meeting host to spread between all participants. In our webapp you can see the video and easily navigate thru it by clicking the transcript below

How we built it

We used Python as a backend, and FLask to create an API server. Frontend was made with React.js and Next.js.

Challenges we ran into

Hard topic to create solution. Getting the video and transcript from meeting- we solved it by using Azure API that generates transcript from audio.

Accomplishments that we're proud of

Working product (frontend, backend) Teamwork Solutions

What we learned

Working in teams. Importance for leader in team. Python, working with transcripts, analyzing speech, API, e-mail sending, devops

What's next for Horoscrypt

Voice volume, loudness and intensity analysis implementation in webapp, Improving algorithms

how it works?

So, that’s it. We are the Horoscript team and we are exceptionally proud of, that we managed to build a wholistic webapp, that also provides a great UX experience for taking notes and keeping better track of your meetings

What is the most important value for meeting participant to gain?

• get down the exact thoughts, said in teams. Features we can provide:
• Summarized meeting transcription
• AI generated To-Do List … put together in the form of follow-up email search for keywords: AI will extract keywords AI as natural language procesing based on analysis we can provide keywords we can search entities (difference types such as next meeting dates, topics came up conversation, people in the conversation, sentimental data to show, how well the meeting went our own transcript, bc teams do not provide api for transcription, so we went with azure api.

oAzure does: Speech API api (for text caption) Text analytics API cognitive cloud (to creat summarizations) flask api python in the backend, flask was used to create api server front in Next JS + React api calls from frontend to backend server

Built With

• azure
• flask
• javascript
• next.js
• pip
• python
• react.js
• requests
• styled-components