diff --git a/Procfile.prod b/Procfile.prod
new file mode 100644
index 000000000..680d5163c
--- /dev/null
+++ b/Procfile.prod
@@ -0,0 +1,5 @@
+marc_bodoni_prod_indexer: JRUBY_OPTS=-J-Xmx8192m /usr/local/rvm/bin/rvm jruby-9.4.1.0 do bundle exec traject -c ./lib/traject/config/sirsi_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_marc_bodoni_prod_indexer.log -s processing_thread_pool=18 -s kafka.topic=marc_bodoni -s kafka.consumer_group_id=traject_marc_bodoni_prod -s solr.url=http://sul-solr.stanford.edu/solr/searchworks-prod -s reserves_path=/data/sirsi/bodoni/crez
+marc_morison_prod_indexer: JRUBY_OPTS=-J-Xmx8192m /usr/local/rvm/bin/rvm jruby-9.4.1.0 do bundle exec traject -c ./lib/traject/config/sirsi_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_marc_morison_prod_indexer.log -s processing_thread_pool=4 -s kafka.topic=marc_morison -s kafka.consumer_group_id=traject_marc_morison_prod -s solr.url=http://sul-solr.stanford.edu/solr/searchworks-morison
+sdr_prod_indexer_catchup: /usr/local/rvm/bin/rvm ruby-3.1.2 do bundle exec traject -c ./lib/traject/config/sdr_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_sdr_prod_indexer_catchup.log -s kafka.topic=purl_fetcher_prod -s kafka.consumer_group_id=traject_catchup -s solr.url=http://sul-solr.stanford.edu/solr/searchworks-prod
+sdr_preview_indexer: /usr/local/rvm/bin/rvm ruby-3.1.2 do bundle exec traject -c ./lib/traject/config/sdr_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_sdr_preview_indexer.log -s kafka.topic=purl_fetcher_prod -s kafka.consumer_group_id=traject_sdr_preview_prod -s purl_fetcher.target=SearchWorksPreview -s purl_fetcher.skip_catkey=false -s solr.url=http://sul-solr.stanford.edu/solr/sw-preview-prod
+earthworks_prod_indexer: /usr/local/rvm/bin/rvm ruby-3.1.2 do bundle exec traject -c ./lib/traject/config/geo_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_earthworks-prod-indexer.log -s kafka.topic=purl_fetcher_prod -s kafka.consumer_group_id=earthworks-prod-indexer -s solr.url=http://sul-solr.stanford.edu/solr/earthworks-prod
\ No newline at end of file
diff --git a/Procfile.stage b/Procfile.stage
new file mode 100644
index 000000000..f23704a5f
--- /dev/null
+++ b/Procfile.stage
@@ -0,0 +1,6 @@
+marc_bodoni_dev_indexer: JRUBY_OPTS=-J-Xmx8192m /usr/local/rvm/bin/rvm jruby-9.4.1.0 do bundle exec traject -c ./lib/traject/config/sirsi_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_marc_bodoni_dev_indexer.log -s processing_thread_pool=2 -s kafka.topic=marc_bodoni -s kafka.consumer_group_id=traject_marc_bodoni_dev -s solr.url=http://sul-solr.stanford.edu/solr/searchworks-dev -s reserves_path=/data/sirsi/bodoni/crez
+marc_morison_dev_indexer: JRUBY_OPTS=-J-Xmx8192m /usr/local/rvm/bin/rvm jruby-9.4.1.0 do bundle exec traject -c ./lib/traject/config/sirsi_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_marc_morison_dev_indexer.log -s processing_thread_pool=2 -s kafka.topic=marc_morison -s kafka.consumer_group_id=traject_marc_morison_dev -s solr.url=http://sul-solr.stanford.edu/solr/searchworks-morison-dev -s reserves_path=/data/sirsi/morison/crez
+sw_dev_indexer: /usr/local/rvm/bin/rvm ruby-3.1.2 do bundle exec traject -c ./lib/traject/config/sdr_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_sw_dev_indexer.log -s kafka.topic=purl_fetcher_prod -s solr.url=http://sul-solr.stanford.edu/solr/searchworks-dev
+sw_preview_stage_indexer: /usr/local/rvm/bin/rvm ruby-3.1.2 do bundle exec traject -c ./lib/traject/config/sdr_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_sw_preview_stage_indexer.log -s kafka.topic=purl_fetcher_stage -s kafka.consumer_group_id=traject_purl_fetcher_stage_sw_preview -s purl_fetcher.target=SearchWorksPreview -s purl_fetcher.skip_catkey=false -s purl.url=https://sul-purl-test.stanford.edu -s solr.url=http://sul-solr.stanford.edu/solr/sw-preview-stage
+earthworks_stage_indexer: /usr/local/rvm/bin/rvm ruby-3.1.2 do bundle exec traject -c ./lib/traject/config/geo_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_earthworks-stage-indexer.log -s kafka.topic=purl_fetcher_stage -s kafka.consumer_group_id=earthworks-stage-indexer -s purl.url=https://sul-purl-stage.stanford.edu -s stacks.url=https://sul-stacks-stage.stanford.edu -s geoserver.pub_url=https://earthworks-geoserver-stage-b.stanford.edu/geoserver -s geoserver.stan_url=https://earthworks-geoserver-stage-a.stanford.edu/geoserver -s solr.url=http://sul-solr.stanford.edu/solr/earthworks-stage
+folio_dev_indexer: /usr/local/rvm/bin/rvm ruby-3.1.2 do bundle exec traject -c ./lib/traject/config/folio_config.rb -s solr_writer.max_skipped=-1 -s log.level=debug -s log.file=log/traject_folio_dev_indexer.log -s kafka.topic=folio_test -s kafka.consumer_group_id=traject_folio_dev -s reader_class_name=Traject::KafkaFolioReader -s solr.url=http://sul-solr.stanford.edu/solr/searchworks-folio-dev
\ No newline at end of file
diff --git a/README.md b/README.md
index 4bfbd4f7d..e5c0a5617 100644
--- a/README.md
+++ b/README.md
@@ -29,11 +29,22 @@ bundle exec rake
 ```
 note that some integration tests may hit a live server, for which you may need to be on the Stanford VPN.
 
+## Building services
+For development we can use Foreman to run a procfile, but on a deployed machine, we export the rules to systemd:
+```
+foreman export -a traject -f Procfile.stage --formation marc_bodoni_dev_indexer=1,marc_morison_dev_indexer=1,folio_dev_indexer=8,sw_dev_indexer=2,sw_preview_stage_indexer=2,earthworks_stage_indexer=1 systemd ~/service_templates
+sudo cp /opt/app/indexer/service_templates/* /usr/lib/systemd/system/
+
+sudo systemctl enable traject.target
+sudo systemctl start traject.target
+```
+
 ## indexing data
 indexing is a multi-step process:
 1. an extractor process publishes data to be indexed to a [kafka](https://kafka.apache.org/) topic
-2. a daemon run by [eye](https://github.com/kostya/eye) consumes data from the kafka topic and invokes traject
+2. Systemd runs the various traject services
 3. traject uses a given configuration to index the data into a solr collection
+
 ### publishing data to kafka
 extractor processes are written as ruby scripts in `script/` and usually invoked by shell scripts located in the same directory. they make use of traject extractor classes stored in `lib/traject/extractors/`, which use the ruby kafka client to publish data to a kafka topic using the pattern:
 ```ruby