Security
Boring, predictable security for AI agents.
Rails Agent is built by a small team of long-time Rubyists. We treat security like a Rails migration — reversible, versioned, and reviewable.
Where your data lives
Every workspace maps to an isolated Postgres schema and a dedicated Vercel project inside our infrastructure. Your agent code lives in your GitHub repo — we only cache builds.
Secrets & API keys
- Model provider keys encrypted with AES-256-GCM at rest
- Per-environment scoping (development, staging, production)
- Optional bring-your-own-KMS for Business plan
- Rotated with a single `rails agent:rotate_key` command
Traces & PII
Traces are stored for 30 days by default. You can redact fields with the `pii:` DSL or disable persistence entirely. Extended retention is available on Enterprise plans.
Compliance
- SOC 2 Type II — in progress (target Q2 2026)
- GDPR-compliant data processing addendum available
- HIPAA BAA on Business plan
- Public status page & post-mortems