Scrape Request

In order for a client to create a scrape request, it must first determine the HTTP URL for the scrape endpoint. This is done by locating the string announce in the path section [3] of the announce URL and replacing it with the string scrape. Performing a scrape request to URLs that are not determined by this method are outside of the scope of this specification.

A scrape request contains no HTTP body, but instead uses HTTP query parameters [3] to encode its key-value pairs. There is only one key, the info_hash key specified in BEP 003 [1]. This key can be appended multiple times with different values for collecting data about multiple swarms in one request.

Scrape Response

The response to a successful request is a bencoded dictionary containing one key-value pair: the key files with the value being a dictionary of the 20-byte string representation of an infohash paired with a dictionary of swarm metadata. The fields found in the swarm metadata dictionary are as follows:

complete

The number of active peers that have completed downloading.

incomplete

The number of active peers that have not completed downloading.

downloaded

The number of peers that have ever completed downloading.

The response to an unsuccessful request is a bencoded dictionary with the key failure_reason and a string value as described in BEP 003 [1].

Example

Request:

http://tracker/scrape?info_hash=xxxxxxxxxxxxxxxxxxxx&info_hash=yyyyyyyyyyyyyyyyyyyy

Response:

d5:filesd20:xxxxxxxxxxxxxxxxxxxxd8:completei11e10:downloadedi13772e10:incompletei19e
20:yyyyyyyyyyyyyyyyyyyyd8:completei21e10:downloadedi206e10:incompletei20eee

Response (JSON encoded for readability):

{
  'files': {
    'xxxxxxxxxxxxxxxxxxxx': {'complete': 11, 'downloaded': 13772, 'incomplete': 19},
    'yyyyyyyyyyyyyyyyyyyy': {'complete': 21, 'downloaded': 206, 'incomplete': 20}
  }
}