Inspiration

Our inspiration for choosing this challenge was the great potential of automated transcription solutions and their future impact.

What it does

The bot creates Microsoft Teams summarization text and shares them with meeting participants.

How we built it

Firstly, we brainstormed and planned the work. Then, we defined minimum required thresholds. Finally, we shared responsibilities within our team and started to work.

Challenges we ran into

Bot deployment to Azure environment, testing Microsoft Teams bot locally, getting the transcript of the video.

Accomplishments that we're proud of

We managed to write a code that gets video, extracts audio, gets transcription from it, summarizes it using an NLP library, and send it to the participants. We managed to write a plugin that can detect changes on the chat and act upon them. This can later be integrated into the other code to be executed automatically when a meeting ends. We learned many new things!

What we learned

We learned Azure environment, building an automated bot, and integrations among different platforms, utilizing Microsoft Graph API, text summarization methods, video to audio conversion, and speech-to-text conversion.

What's next for Codex Teams Bot

Seams will provide businesses with a tool for online meeting transcription bots worldwide. As further development, we can integrate our text summary generating script to the Teams bot to generate text summary automatically when a meeting ends. Text summarization methods can be improved by training a custom model. We can also use a better Speech To Text API to improve its accuracy.

Built With

Share this project:

Updates