Platform
Build, test, deploy, and monitor Rails AI agents
Rails Agent is organized around the production lifecycle—not a random feature list. Start by building agents in your Rails app, prove behavior with evaluations, deploy on the hosted runtime, then monitor traces, spend, and policy.
Updated 2026-07-28 · Tiny Bubble Company
Build
Define agents under app/agents/, attach tools and knowledge, and scaffold from Playbooks.
Test
Run sandbox traffic and evaluation suites before you promote an agent to production.
Deploy
Promote to the hosted runtime, connect channels, and keep long work on background execution.
Monitor
Watch traces and costs, enforce budgets and guardrails, and require approvals for risky actions.
Frequently asked questions
What does “full-stack agentic platform” mean?
It means Rails Agent covers the agent lifecycle—build, test, deploy, and monitor—plus channels, playbooks, memory, and guardrails—rather than only providing a model client or a thin agent DSL.
Where should a new team start?
Start with Get started and the build guide, scaffold from a Playbook, add evaluations, then deploy and turn on observability.