We ingest massive amounts of data, from about tens of thousands of Internet radio stations, 24/7. We store this data and crunch away, making sense of it all, using sophisticated mathematical methods from the fields of machine learning and collaborative filtering. Our proprietary methods help our Llamas compute similarities between tracks, artists and stations so that we can recommend the stations you are likely to want to listen to.

Our database is working in the background to master all the various spellings of every one of the over 7 million tracks and many thousands of artists. Try spelling the name a little differently and see if that helps!

We are working with playlist data reported by the stations themselves and it is often imperfect. We are doing our best to reformat it in the most readable way.

You can search by country and city. Results are sorted by how recently we have collected data from the stations so you get the freshest plays first!

The list of results you get from Radio One Llama is carefully ranked and ordered to give you the stations that are most relevant to your search, taking into account how often a station plays your favorite artist and similar artists.

The world of Internet radio and music evolves rapidly. Instead of deciding on a fixed genre list, like traditional systems, our genres are computed dynamically, based on what's currently getting airplay on stations.

Although most stations give themselves a genre label, which we do use, Radio One Llama has a proprietary method to identify a station's genre by analyzing all the artists being played on it, to deduce its genre in an objective way. Hence, if an artist is particularly popular in region (but not necessarily from there) they might be labelled with that tag (e.g. Turkish or Croatian).

Here are the facts on who has been listening to online radio in the last month:

• Estimated 49 million Americans
• 20% of U.S. population 12+

Those who have listened to online radio in the last week:

• Estimated 29 million Americans
• 11% of U.S. population 12+
• 16% of Adults 18-34
• 14% of Adults 18-49

Arbitron Radio Listening Report 2007

• Our experts estimate that there are over 100,000 Internet Radio Stations streaming worldwide. We are constantly "discovering" new ones.
• There are over 24 thousand unique Shoutcast stations that we have indexed. Radio One Llama "crawls" 10,000-20,000 Shoutcast stations per day and is always discovering new ones, and updating what it knows about the rest, once about every 9 minutes to be exact. Stations may duplicate themselves for accommodating larger audiences and multiple bit rates.
• There are over 500,000 Shoutcast listeners at a given time.
• One Llama Media has "listened" to over a million tracks.

We do our best to include as many artists as possible in our search. If your artist is not found it may mean that the name is misspelled and we can't interpret the spelling, or that the artist is not showing up on the playlists of the stations we are currently searching. If your artist is missing, please shoot us the artist's name and we will find them!

We extract artist and title information that the radio stations give us. We combine this data in unique ways to extract the best information from each one.

Half the tracks played on Shoutcast Stations have no metadata so we can't find the associated information. Although we have information on over 10.5 million tracks, it is impossible to have the data on everything!

Could be a slow connection, a firewall or the station just might be overloaded with listeners for the bandwidth that they have.

The Radio One Llama Station Rank algorithm ranks search results by combining playlist data with artist-to-artist and station to station similarity. We also take into account how relevant a station is to the artist you are searching for and how recently it played the relevant tracks. Radio One Llama uses Machine Learning techniques to deduce similarity.

Although most stations give themselves a genre label, Radio One Llama uses its own proprietary method to identify a station's genres and then compares that to all the artists being played on a station to deduce its genre in an objective way.

Radio One Llama uses very scalable technology with rapid, high capacity learning models that can ingest tens of thousands of click-streams and millions of objects per second.

One Llama Media is a software development start-up born out of the University of Illinois at Champagne/Urbana. We specialize in a mathematically advanced area of computer science called Machine Learning, which we have applied to collaborative filtering and the analysis and recommendation of music. The company was founded by a group of computer scientists and researchers at the University of IL. and the National Center for Supercomputing Applications and based on work that was being done in the fields of advanced data mining and machine learning.