This project develops an autonomous AI agent that generates engaging speeches by automatically retrieving and processing detailed company and country information. The system combines data from pre-prepared JSON documents (extracted from Wikipedia) with live internet data from the Tavily API to create high-quality, up-to-date speeches—all within an interactive Jupyter Notebook environment.
- Automated Data Retrieval:
- Extracts data from JSON files containing company and country information.
- Fetches real-time data using the Tavily API.
- Advanced Text Generation:
- Utilizes GPT-4 to generate, review, and refine speech drafts.
- Efficient Data Handling:
- Employs Pinecone for vector-based similarity searches on pre-prepared documents.
- Iterative Workflow:
- Uses a directed workflow graph (via LangGraph) to manage planning, content generation, review, and iterative revisions.
- Modular Architecture:
- Designed for scalability and integration with additional expert agents and future enhancements.
-
Clone the Repository:
git clone https://github.com/HabtamuFeyera/InfoSpeakAI.git cd InfoSpeakAI -
Install Dependencies: Install the required packages using pip:
pip install -r requirements.txt
-
Configuration:
- Set up your API keys for GPT-4, Pinecone, and Tavily.
-
Launch Jupyter Notebook: Start the notebook server:
jupyter notebook
Open the main notebook file in your browser.
-
Run the Notebook:
- Execute the notebook cells sequentially to initialize the environment, process data, and generate the speech.
- The notebook is structured to guide you through data extraction, content generation, and iterative refinements.
- Expand multi-agent collaboration with additional expert agents.
- Integrate dynamic user feedback for continuous improvement.
- Enhance data processing algorithms for even more accurate information retrieval.