diff --git a/.cfignore b/.cfignore
new file mode 120000
index 00000000..3e4e48b0
--- /dev/null
+++ b/.cfignore
@@ -0,0 +1 @@
+.gitignore
\ No newline at end of file
diff --git a/.gitignore b/.gitignore
index a395c345..ca37456c 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,7 +9,6 @@ db/*.db
 .*.sw*
 .env
 .*.env
-.cfignore
 .vagrant
 .idea/
 *profile*
diff --git a/create-space.sh b/create-space.sh
new file mode 100755
index 00000000..ea3c9a66
--- /dev/null
+++ b/create-space.sh
@@ -0,0 +1,51 @@
+#!/bin/bash
+set -e
+
+if [ ! -n "$1" ]; then
+  echo "Usage: From inside open-data-maker directory..."
+  echo "  create-space.sh spacename"
+  exit 1
+fi
+echo "Creating space: $1"
+SPACE=$1
+
+# Creating the indexing space
+cf create-space ${SPACE} -o ed
+
+echo "# When creating the space, the user which creates it is a SpaceManager"
+echo "# Add developers using: cf set-space-role USERNAME ed ${SPACE} SpaceDeveloper"
+
+# Target the space
+cf target -o ed -s ${SPACE}
+
+# Create a bservice instance used by snapshot tools
+cf create-service s3 basic data-files
+
+
+#cf create-user-provided-service bservice -p '{"BSERVICE_ACCESS_KEY":"YOUR_S3_ACCESS_KEY", "BSERVICE_SECRET_KEY": "YOUR_S3_SECRET_KEY", "BSERVICE_BUCKET": "YOUR_S3_BUCKET"}'
+
+# Create backup service
+echo "# To create a backup service in this space run:"
+echo "# cf create-service s3 basic backup"
+
+# Create the ElasticSearch service
+cf create-service elasticsearch-swarm-1.7.5 3x eservice
+
+echo "Creating the API server by pushing the ccapi-${SPACE} app:"
+cf push -f manifest-${SPACE}.yml
+echo "By default the app will use the data-files bucket, leaving DATA_PATH env blank"
+
+echo "# For data archive / downloads, these are served via a S3 proxy"
+echo "# The /downloads path is redirected via CloudFront to"
+echo "# ed-public-download.apps.cloud.gov which is in the production space"
+echo "# To create additional S3 proxies: https://github.com/18F/cg-s3-proxy"
+
+echo "TODO: how to put files in the bucket"
+
+echo "# now you need to index"
+echo "cf-ssh -f manifest-${SPACE}.yml --verbose"
+echo "# wait several minutes for this to connect"
+echo "echo $DATA_PATH"
+echo "# it should be blank, meaning you will get default cities data"
+echo "rake import"
+echo "# when this is done, go to https://ccapi-${SPACE}.18f.gov and explore"
diff --git a/lib/data_magic.rb b/lib/data_magic.rb
index 1000d5cf..b48879ca 100644
--- a/lib/data_magic.rb
+++ b/lib/data_magic.rb
@@ -41,8 +41,11 @@ class InvalidDictionary < StandardError
 
   def self.s3
     if @s3.nil?
-      if ENV['VCAP_APPLICATION']
-        s3cred = ::CF::App::Credentials.find_by_service_name(ENV['s3_bucket_service'] || 'bservice')
+      if ENV['VCAP_APPLICATION']  # in Cloud Foundry
+        cfcred = ::CF::App::Credentials.find_by_service_name(ENV['s3_bucket_service'] || 'data-files')
+        s3cred = {'access_key'=> cfcred['access_key_id'], 'secret_key' => cfcred['secret_access_key']}
+        bucket_name = cfcred['bucket']
+        # TODO: default DATA_PATH to s3://{bucket_name}
       else
         s3cred = {'access_key'=>  ENV['s3_access_key'], 'secret_key' => ENV['s3_secret_key']}
       end
diff --git a/manifest-testing.yml b/manifest-testing.yml
new file mode 100644
index 00000000..94f74014
--- /dev/null
+++ b/manifest-testing.yml
@@ -0,0 +1,16 @@
+---
+applications:
+-  name: ccapi
+   host: ccapi-testing
+   domain: 18f.gov
+   command: bundle exec puma -C ./config/puma.rb
+   instances: 1
+   memory: 1G
+   services:
+   - data-files
+   - eservice
+   env:
+     MAX_THREADS: 5
+     WEB_CONCURRENCY: 1
+     INDEX_APP: enable
+     NPROCS: 1
diff --git a/script/s3config.rb b/script/s3config.rb
index d54fb6c3..c0d993dd 100644
--- a/script/s3config.rb
+++ b/script/s3config.rb
@@ -4,6 +4,7 @@
 #    @s3 = ::Aws::S3::Client.new
 
 require 'dotenv'
+require 'json'
 
 branch = `echo $(git symbolic-ref --short HEAD)`.chomp
 
@@ -12,26 +13,34 @@
   puts "using APP_ENV from environment #{APP_ENV}"
 else
   case branch
-    when "master"
-      APP_ENV = "production"
-    when "staging"
-      APP_ENV = "staging"
-    else
-      puts "not on master or staging branch lets use dev"
-      APP_ENV = "dev"
+  when "master"
+    APP_ENV = "production"
+  when "staging"
+    APP_ENV = "staging"
+  else
+    puts "not on master or staging branch lets use dev"
+    APP_ENV = "dev"  # FIXME: shouldn't the APP_ENV be testing?
   end
 end
 
-Dotenv.load(
-  File.expand_path("../../.#{APP_ENV}.env", __FILE__),
-  File.expand_path("../../.env",  __FILE__))
+cf_credentials = `cf target -o ed -s #{APP_ENV} && echo "$(cf env ccapi)" | tail -n +5 |  sed -n -e :a -e '1,10!{P;N;D;};N;ba'`
+cf_json_str = cf_credentials.gsub("\n", '').gsub(/^[^{]+{/, '{').gsub(/{\s+"VCAP_APPLICATION".+$/, '')
+
+cf_json = JSON.parse(cf_json_str)
+cf_data_files = cf_json['VCAP_SERVICES']['s3'].detect {|j| j['name'] == 'data-files'}
+
+fail "Unable to find data-files configuration" if cf_data_files.nil?
+
+ENV['CF_CREDENTIALS'] = cf_json_str
+ENV['AWS_ACCESS_KEY_ID'] = cf_data_files['credentials']['access_key_id']
+ENV['AWS_SECRET_ACCESS_KEY'] = cf_data_files['credentials']['secret_access_key']
+ENV['BUCKET_NAME'] = cf_data_files['credentials']['bucket']
 
 require 'aws-sdk'
 puts "app env: #{APP_ENV}"
-puts "bucket name: #{ENV['s3_bucket']}"
-
+puts "bucket name: #{ENV['BUCKET_NAME']}"
 
-s3cred = {'access_key'=>  ENV['s3_access_key'], 'secret_key' => ENV['s3_secret_key']}
+s3cred = {'access_key'=>  ENV['AWS_ACCESS_KEY_ID'], 'secret_key' => ENV['AWS_SECRET_ACCESS_KEY']}
 
 ::Aws.config[:credentials] = ::Aws::Credentials.new(s3cred['access_key'], s3cred['secret_key'])
 ::Aws.config[:region] = 'us-east-1'
diff --git a/script/s3pull b/script/s3pull
index 7cb48ec3..e9647f9c 100755
--- a/script/s3pull
+++ b/script/s3pull
@@ -4,7 +4,7 @@ require_relative 's3config.rb'
 
 @s3 = ::Aws::S3::Client.new
 
-bucket = ENV['s3_bucket']
+bucket = ENV['BUCKET_NAME']
 
 dirname = 'real-data'
 unless File.directory?(dirname)
diff --git a/script/s3push b/script/s3push
index 6d300280..c4ed1bee 100755
--- a/script/s3push
+++ b/script/s3push
@@ -5,7 +5,7 @@ require_relative 's3config.rb'
 @s3 = ::Aws::S3::Client.new
 
 dirname = 'real-data'
-bucket_name = ENV['s3_bucket']
+bucket_name = ENV['BUCKET_NAME']
 datayamlpath = File.expand_path("../../#{dirname}/#{bucket_name}.yaml",  __FILE__)
 
 puts "copying #{datayamlpath}"