I stumbled upon your service today. I'm trying to understand how I could use it.
I have some questions then:
I upload a song for analysis, and get the XML result back.
What happens if I want to do the same query a day or a week after.
Do I need to upload the song again?
Do you cache it in some way?
Do I need to cache the query/result?
Is there any other API that I can use to provide a song, rather than uploading it?
Put in another way. Can you read the song from somewhere else than an upload?
Any other music service? Spotify? Last.fm?
alesch: when you upload a track, you get an ID back. You can use that ID to retrieve your analysis for a long time (months, years). Also, if you've lost the ID, you can still short circuit the upload by generating an MD5 of the track and using that as an ID. Most of the client libraries (like pyechonest or the the java echo nest client library) will do this MD5 check automatically.
If there's a URL that points to audio you can use that with the 'url' parameter of the upload call. The url has to point to a publicly available audio file (this means that it won't work with spotify or last.fm urls). Note also that we are just releasing an alpha version of a search_tracks api that will let you search for analysis data by track and artist name. Read more about it here: