Inferra is an open-source platform for hosting, sharing, and running pre-trained Machine Learning, Deep Learning, and NLP models.
It provides a simple and interactive UI to explore models, make predictions, and deploy them with ease, making model inference accessible to teams, researchers, and enthusiasts alike.

We are a group of students from the Faculty of Engineering, Cairo University, working to make AI models easy to use, share, and deploy.

• Central Model Hub – Store and manage all your models in one convenient place.
• Easy Inference – Run predictions from the UI with minimal setup.
• Deployment Ready – Seamlessly deploy models for production or research use.
• Framework Agnostic – Compatible with PyTorch and TensorFlow frameworks.
• Community Driven – Share models, ideas, and improvements with the AI community.

To create a platform where AI models can be easily hosted, explored, and used, empowering researchers, developers, and enthusiasts to leverage AI without the overhead of setup or integration.

If you just want to try Inferra without setting up the development environment, you can run the app directly here:

👉 Inferra Live App

We welcome contributions from everyone — whether you want to improve the platform, add new models, or help with documentation. Here’s how you can get started:

1. Fork the repository and create your branch from main.

2. Clone your fork and set up the development environment:

   git clone https://github.com/your-username/inferra.git
   cd inferra
   pip install -r requirements.txt

3. Create a new branch for your feature or bugfix:

   git checkout -b my-feature

4. Make your changes, following these guidelines:

   • Add new model architectures in src/models.
   • If you create new layers, add them under src/layers/.
   • To create your own app in Inferra, make a new directory inside app/apps/ with the desired app name (e.g., app/apps/my_app/); the directory name will be used to generate the app’s display name in the UI (e.g., my_app → My App). This directory must contain an app.py file—without it, the app will not be recognized or runnable by Inferra.

      app/
      └─ apps/
        └─ my_app/
          └─ app.py
     

   • Upload trained model weights to cloud storage (Hugging Face, Google Drive, AWS S3, etc.) — do not include large model files in the repo.

5. Commit and push your changes:

   pre-commit run --all-files --hook-stage manual
   git add .
   git commit -m "Describe your changes"
   git push origin my-feature

6. Open a Pull Request on GitHub and describe your changes clearly.

• Add new pre-trained models or layers.
• Improve the web interface and UX.
• Help with documentation and tutorials.
• Test the platform and provide feedback.

Happy coding! Join the Inferra community and help make AI accessible for everyone!