I think this is a novel idea and I congratulate you folks.
One thing of a concern I have is negative marketing. What if a group of people have negative interest in a business and flood your website with negative-but baseless-comments? This has been plaguing other internet review sites and companies are deliberately doing it to make money out of "cleaning the reviews." I am curious if you had any plans to counter that or will just let market correct itself.
Good luck.

 This will work if you have proper monitoring mechanism in place.  Take your cue from Amazon merchant rating system or Yelp.  Trip Advisor came under fire lately because its ratings were getting grossly manipulated.

 I think this is a great idea, has a great potential. Congratulate guys, and I am sure this will pick up. baru nepali ma pani lekhna milne garnu parne. For this site to take off, you know you need the review on small restaurants in Itahari or Tandi too, not merely a few glitzy restaurants of KTM.

 Stiffler and Nina_windsor,
 
   Both of you are absolutely correct, i.e. how to monitor the review site so that it is not rigged by spammers and businesses voting themselves?

In fact there are several problems that we already face;

1) Business placing 5 star reviews for their own business
2) Business owners lowering the reviews of rival businesses
3) Spamming the business reviews with non-sense information

We have already banned a few people and deleted many self-review businesses :)

We are doing various things to address this. 

First, is the crowdsourcing approach. Any review can be flagged by other users so if comments look deceptively good or is a spam it can be flagged. We manually go through such flagged comments at http://www.parakhi.com

Second, we are using various automatic methods, including keeping track of who is saying what about which businesses and finding correlations to predict if it's the business owner who is reviewing their own businesses.  The second approach is more advanced Machine Learning approach that we are building in-house. This module actually not only finds such non-sense reviews but it also predict user choices by giving recommendations to the users on restaurants or businesses they might like :)

Pire, Currently we have information collection people going to each business and taking pictures and making sure collected information is accurate. It is very expensive process but we are doing it to make sure information is accurate. Hence, going to Ithari and Tadi to collect information is not possible for now :( 

I like your approach, very professional and logical. Keep it going, we will support it.