Scrapyd with Selenium Spider

selenium_spider.py

from selenium import selenium
from scrapy.spider import BaseSpider
from scrapy.http import Request
import time
import lxml.html

class SeleniumSprider(BaseSpider):
    name = "selenium"
    allowed_domains = ['selenium.com']
    start_urls = ["http://localhost"]
    
    def __init__(self,  **kwargs):
        print kwargs
        self.sel = selenium("localhost", 4444, "*firefox","http://selenium.com/")
        self.sel.start()
    
    def parse(self, response):
        sel = self.sel
        sel.open("/index.aspx")
        sel.click("id=radioButton1")
        sel.select("genderOpt", "value=male")
        sel.type("nameTxt", "irfani")
        sel.click("link=Submit")
        time.sleep(1) #wait a second for page to load
        root = lxml.html.fromstring(sel.get_html_source())

Quick question, are the allowed domains the same one you put for the definition of the sel object?
Why is the start url localhost? does it mean localhost from selenium.com or the localhost of your machine? (for some strange reason)