This app is designed as a learning aid for anyone seeking to become familiar with Generative AI
apps and agents. It's based on the user interface in Microsoft Foundry portal, but does not use
any Azure cloud services.
The app uses Microsoft's Phi 4-mini on wllama (works on any device
without requiring a GPU). You can switch between Phi 4-mini and None (Wikipedia) using the Model
dropdown.
The app also makes use of the MobileNetV3 model and the TensorFlow.js framework to implement
image classification.
If AI mode responses are too slow or become unresponsive on older or lower-spec
computers, switch the Model dropdown to None as a last-resort option. This
mode skips local language model inference and uses a lightweight fallback experience.
Known issues
- The initial download of the Microsoft Phi 4-mini model may take a few minutes - particularly on
low-bandwidth connections. Subsequent downloads should be quicker.
- Phi 4-mini response times may be slower on lower-spec devices since the model runs
locally through WebAssembly. Use the Model dropdown to switch to None if needed.
- Image analysis is based on MobileNetV3, which is a general-purpose image classification
model. It may not accurately identify all images or provide detailed descriptions.
- Voice mode depends on browser-support for speech recognition. A
fallback offline speech-to-text model will be used in the event of a speech recognition
failure.
- Voice mode performance may vary based on the device's microphone quality and ambient noise
levels. For best results, use a good quality microphone in a quiet environment.