The CV to HTML/CSS/JS-Portfolio Website Generator is an AI-powered application designed to transform your CV (in PDF or DOCX format) into a beautiful and responsive static HTML portfolio. Utilizing advanced tools such as CrewAI and Google Gemini, this project streamlines the process of showcasing your skills and experiences in a professional manner.
Try now: https://huggingface.co/spaces/Sobit/cv-2-website
Full tutorial: https://medium.com/@nepalsobit1
- CV Analysis: Analyzes your CV to extract key skills, experiences, and accomplishments.
- Portfolio Generation: Generates a fully responsive HTML/CSS/JS portfolio website based on the analysis.
- Modern Design: Employs a clean layout with sections for skills, projects, experiences, certifications, publications, and contact details.
- Easy Deployment: The output is a complete HTML document ready for deployment.
- CrewAI: Tool used for creating AI agent for CV analysis and portfolio generation.
- Google Gemini: Language model for understanding and generating content.
- Gradio: Framework for creating user-friendly web interfaces.
- pdfplumber: For extracting text from PDF files.
- python-docx: For reading DOCX files.
Make sure you have the following installed:
- Python 3.x
- pip (Python package installer)
- Download the notebook file and run all the cells to host on your own.
- Clone this repo.
git clone https://github.com/sobit-nep/cv-to-portfolio-crewai-agent.git
- Install all the required dependencies.
pip install -r requirements.txt
- Start the gradio app by running,
python gradio_app.py
-
Click on the Upload button to upload and generate your portfolio.
-
The preview webpage for generated code will be displayed, and you can download the complete code by clicking on the Download Code button.
Contributions are welcome! If you have suggestions or improvements, feel free to create an issue or submit a pull request.
This project is licensed under the MIT License - see the LICENSE file for details.
- Thanks to CrewAI and Google Gemini for their powerful tools that made this project possible.
- P.S. As I've used the freely available Gemini model here, the output isn't that convincing. I tried OpenAI's gpt 4-o and the result was 100x better in terms of responsiveness and UI.