Data Flow and Privacy

A technical boundary map of what VibeRails executes locally and where model requests go.

Execution boundary

VibeRails runs on your machine. It reads local repositories you explicitly add, then orchestrates configured CLI backends for review/fix execution.

Model request path

When a review or fix session runs, request payloads are sent by the configured CLI backend to your selected model endpoint. That endpoint can be:

• Cloud provider APIs under your account
• Local model servers on your hardware
• Private network endpoints in controlled environments

What VibeRails does not require

• No mandatory VibeRails-hosted source-code storage
• No always-on VibeRails cloud relay
• No proprietary model dependency

Local and air-gapped references

• Local AI code review setup
• Air-gapped deployment options
• Trust center overview