AI Mixtape Generator
What is Natural Language Playlist and how does it work?
Natural Language Playlist is an AI-powered tool that generates playlists using language-powered search. It focuses on musical and cultural features such as genre, lyrical meaning, sonics, and vibes to craft playlists. You can log in with Spotify to have playlists created directly on your Spotify account, or continue as a guest.
How do I login with Spotify and what happens when I do?
Logging in with Spotify generates playlists on your Spotify account and helps improve the AI algorithm over time as it gains insights into your musical preferences. There is also an option to continue as a guest.
How should I describe emotional themes and moods to get the best results?
To create playlists with specific emotional themes, use clear, descriptive language that highlights the mood or vibe you want. For example, phrase requests like "uplifting and energetic mornings" or "melancholic vibes for rainy days." The model responds best to explicit emotions and atmospheres.
Can I use artist names or track titles in my queries?
You can, but Natural Language Playlist excels at understanding broad musical and cultural themes. For better results, describe the genre, vibe, or lyrical content instead of relying solely on specific artist or track names.
What are the limitations of Natural Language Playlist?
- Limited Catalog: the dataset may not include every song or artist, potentially missing niche or lesser-known tracks.
- Language Sensitivity: the tool’s interpretation may not always perfectly align with your intended mood or theme.
- Dependency on Input: the quality of playlists relies on how clear and specific your language is.
- No User Interaction: the tool does not adapt its recommendations based on ongoing user feedback in real time.
Who created Natural Language Playlist, and what's its background?
Natural Language Playlist was created by Abelardo Riojas, a 23-year-old data science graduate student and self-described “music dude.” The project uses a curated dataset of textual song metadata to build playlists that reflect musical and cultural traits.
How can I follow the project or contact the creator?
You can follow the project on Instagram at @notabelardoriojas. For business or professional inquiries, you can email ariojas@utexas.edu or reach out via LinkedIn.
Can you share examples of how to describe a playlist?
Yes. Examples of playlist descriptions include:
- “Midwest Emo songs to cry in the shower to because my girlfriend broke up with me.”
- “Absolutely unlistenable songs. Music that doesn't even sound like music.”
- You can also describe the musical features, e.g., “Very danceable, positive mood, fast tempo, high energy, songs for lovers.”
