In this paper we propose an unsupervised, domain independent as well as corpus independent approach for automatic keyword extraction. In second part of the paper we have suggested an extension of the approach to extract keyphrases from the document. The approach is general and can be applied to any language. However, we have tested the approach on Hindi language. Our approach combines the information contained in frequency and spatial distribution of a word in order to extract keywords from a document. Our work is especially significant in the light that it has been implemented and tested on Hindi which is a resource poor and underrepresented language.