Website Link: https://meeting-notes-main.vercel.app/

Google Colab Demo: https://colab.research.google.com/drive/1BWbhKEIG9HMczS0sl8ejKHhrm-KAuziT?usp=sharing

Output Continued

Next.js Application

Follow these steps to get started:

Prerequisites

Ensure you have Node.js and npm installed on your machine. You can download them here.

Installation

.

1. Clone the Repository: Clone this repository to your local machine.

   git clone https://github.com/AnshTanwar/meeting-notes-main.git
   cd meeting-notes-main

2. Install Dependencies: Run the following command to install the project dependencies.

        npm install --legacy-peer-deps

3. Set up environment variables: Add the following key-value pair to the .env file: Create a .env file: In the root directory, create a new file named .env.

        ASSEMBLYAI_API_KEY=<your api key>
        FIREBASE_SERVICE_ACCOUNT_KEY_BASE64=<your api key>
        FIREBASE_STORAGE_BUCKET=<your api key>
   
        

4. Start the Development Server: Run the following command to start the development server.

        npm run dev

5. Access the Application: Open your web browser and navigate to http://localhost:3000.

Usage

The application should be accessible at http://localhost:3000.

Multi-Speaker Transcription Approach Using AssemblyAI

1. Setup: Initialize the AssemblyAI API and configure transcription settings with options like speaker_labels, auto_highlights, and entity_detection.

2. Transcription: Use transcribe() to process the audio and generate the transcript with speaker labels.

3. Extract Speaker-Text Pairs: Loop through the transcript’s utterances, store each speaker’s text in a list as [Speaker n, text].

4. Output: The list of speaker-text pairs is ready for analysis.

---

Meeting Analysis Process

This code utilizes multiple prompts to generate a structured analysis of meeting transcripts, including a summary, counterpoints, ideas, and action items. Below are the steps involved:

1. Generate Meeting Summary

   • Objective: Summarizes key outcomes, action items, and discussed steps from the meeting transcript.
   • Input: Meeting transcript (list of speaker-dialogue pairs).
   • Output: A JSON object containing the meeting outcomes, action items, and discussed steps.

2. Extract Counterpoints and Ideas

   • Objective: Identifies counterpoints and proposed ideas discussed but not adopted during the meeting.
   • Input: Meeting transcript (list of speaker-dialogue pairs).
   • Output: A JSON object containing counterpoints and proposed ideas.

3. Assign Actions and Responsibilities

   • Objective: Organizes action items from the meeting summary into structured tasks with responsibilities, deadlines, and importance levels.
   • Input: Meeting transcript (list of speaker-dialogue pairs) and action items (extracted from the meeting summary).
   • Output: A JSON array containing detailed task assignments (description, DRI, consulted individuals, importance, deadline).

4. Main Function

   • Objective: Calls the above functions sequentially to generate a comprehensive meeting analysis.
   • Input: Meeting transcript (list of speaker-dialogue pairs).
   • Output: A dictionary containing the meeting summary, counterpoints and ideas, and action items with responsibilities.

---

Model Interactions

• Each function interacts with the generative model (gemini-1.5-pro-latest) to produce content based on the specified inputs.
• The model configurations, such as temperature and top_p, are adjusted for each task to control the output.

---

Example Usage

analyze_meeting(meeting_transcript_list)