# Managing data
# Job Data
Job Data represents variables that change between every run of an automation. This is also the information that can be passed in an API call to start a job in Prefix.
There are two places you can specify Job Data in the orchestrator: when adding a model and when creating a new job. If the job associated with an automation model in Prefix requires Job Data, you must specify Default Job Data when adding the model to Prefix. This provides Prefix with a test case to run when you start a job manually from the orchestrator. The only requirement for Job Data is it must be valid JSON.
If you wish to create a job that uses data passed from an API call, then you will require Job Data. There's more info on the exact API in Starting Jobs.
# Data Tables
Data Tables hold tabular data that is either imported as a CSV or Excel file, or represent the outputs of a particular job. These can be used to save scraped data, maintain audit trails, or simply store configuration data in a format automations will be able to refer to.
Unlike Job Data, Data Tables contain persisted data that is not necessarily changed between every job run.
# Files
This is fairly self-explanatory, Files in Prefix provide a simple way to organize input and output files to and from automations. Actual data is stored locally on your orchestrator deployment-- if your organization has many hard limits on where and how much data can be stored, you should consider deploying the orchestrator on-premises.
# Vault
The vault is used to manage secret information for automations to consume at runtime. Credentials entered here are generally secure on our encrypted Postgres database, but we also provide support for storage on Hashicorp's Vault.
Vault information is passed to automations as they are run, and allows your team to write scripts that log in to portals/business software on behalf of human operators without worrying about storing passwords in cleartext inside scripts.