# Retention

> MARFI Insight AI retention is plan-based: 90 days rolling on Insight and 12 months on Trajectory, with automated deletion of expired data.

Canonical URL: https://insight.marfi.app/retention

Retention windows are part of privacy, cost control, and governance. MARFI gives each tenant a plan-based retention policy and enforces expiry automatically.

## Audience

Security, legal, IT, and operations teams reviewing data lifecycle.

## Search Intent

Understand MARFI data retention windows for Insight, Trajectory, and Enclave.

## Telemetry analytics with 90-day rolling retention.

Insight retains operational telemetry long enough to support productivity analysis, workflow mining, and recent trend review without keeping data indefinitely.

## Training datasets need a longer runway.

Trajectory includes 12-month retention because AI training and workflow replay often need richer longitudinal context.

## Expiry is automatic and tenant-scoped.

Retention sweeps apply each tenant's configured window and delete expired rows and stored objects. Custom windows are available where the deployment requires them.

## FAQ

### How long does Insight keep data?

Insight includes 90-day rolling retention.

### How long does Trajectory keep data?

Trajectory includes 12-month retention.

## Related Pages

- [Security](https://insight.marfi.app/security): MARFI Insight AI uses tenant isolation, encryption at rest, TLS in transit, plan gates, and scoped capture controls for workplace telemetry.
- [Privacy](https://insight.marfi.app/privacy): MARFI's privacy model uses consent tiers, telemetry-only Insight capture, customer data ownership, and retention windows.
- [Pricing](https://insight.marfi.app/pricing): MARFI Insight AI pricing is per sensor: Insight at $15 per sensor per month, Trajectory at $99, and Enclave as custom self-hosted annual licensing.
- [Self-Hosted Enclave](https://insight.marfi.app/self-hosted-enclave): MARFI Enclave is the self-hosted deployment option for organizations that need workforce telemetry and AI training data inside their own infrastructure.