OnRails
AI agent governance
A runtime that runs AI agents under a company's business rules. Teams write instructions and rules in an onrails.yaml policy, and OnRails records each run and blocks or asks for sign-off on actions that break a rule.
- Rules in an onrails.yaml policy that block an action unless a check has passed
- Person sign-off on actions such as refunds or write-offs
- Records every program, file write and connection an agent makes during a run
- Enforcement at the operating-system level
- Rules drafted by OnRails and reviewed and turned on by the customer
Built for
Companies putting AI agents into production workflows such as billing, payments and renewals