Question 1

How does Moodify detect my mood?

Accepted Answer

Moodify ships three self-trained models. A BERT classifier reads tone and sentiment from the text you type, a wav2vec-style network reads pace + energy from a short voice recording, and an FER + facenet-pytorch pipeline reads your expression from a single front-camera photo.

Question 2

Where do the recommendations come from?

Accepted Answer

Moodify queries Deezer's free, keyless search endpoint with a mood-tuned keyword phrase (optionally biased by your selected genre and history) and returns up to 60 deduplicated tracks per analysis.

Question 3

Is Moodify free?

Accepted Answer

Yes. Moodify is free for personal use. The web client runs on Vercel and the ML inference runs on Modal serverless GPUs.

Question 4

Does Moodify store my photos or voice clips?

Accepted Answer

No. Photos and voice recordings are sent to the inference service for a single classification call and are never persisted server-side.