In order to access the corpora indexed with Corpus Work Bench (cwb),
you don't necessarily need the Korp interface. Each corpus can be accessed
directly using cqp (Corpus Query Processor).

There is a number of documents describing the query language.
With the keywords "cqp", "query language", "tutorial" it is possible
to retrive the most relevant ones.

An entry point for to use cqp directly is here:

http://cwb.sourceforge.net/files/CQP_Tutorial/

The most relevant part of the tutorial starts at this node:
http://cwb.sourceforge.net/files/CQP_Tutorial/node7.html

The cwb is installed on the gtweb server.

A typical working session:

1. log in to gtweb.uit.no

2. start cqp

> cqp -e

3. list all corpora

[no corpus]> show corpora;

4. choose a corpus (use TAB for completion)

[no corpus]> SME_AVVIR_20140318;
SME_AVVIR_20140318> 

5. search (e.g., for a specific lemma)

SME_AVVIR_20140318> [lemma = "mannat"];

6. save your results to an external file

SME_AVVIR_20140318> cat > "my_text.txt"

A file with the results is then created in the current directory.

7. close session 

SME_AVVIR_20140318> exit;

Done. Share and enjoy!