This day I took on the challenge of setting up a Amazon Micro (Free Tier) machine to run a simple web service for OCR using Tesseract (http://code.google.com/p/tesseract-ocr/).
There is a default web service setup with params:
img: uri to the jpg image that you want to transform
callback: used for crossdomain services
format: deafults to json but you can also supply txt to get just raw text
To build a machine you can follow these steps or if you are interested in the image let us know and I will contact you with more information.