Juin 2026

June 8, 2026

PII masking scope for in-flight detection

The in-flight PII guardrail now supports a configurable masking scope. In addition to detecting PII in text inputs, you can now apply PII masking to files retrieved by agents at runtime and passed as inputs to LLMs.

For details, refer to Settings for AI Trust Layer Policies.

Correction, June 8, 2026: This feature was released on June 5, 2026, but was not included in the release notes at the time.

Unified outbound IP ranges — transition begins

Automation Ops outbound IP ranges are being unified. The new unified outbound IP ranges apply per regions (for example, United States, Europe, etc.) and are the same for all UiPath services within a region.

For the full list of affected services, refer to the June 8, 2026 Automation Cloud release notes.

The current Automation Ops outbound IP ranges will remain active until September 16, 2026. Starting September 16, they will be progressively phased out as the unified IP ranges take effect.

What you need to do:

1. Add the new unified IP ranges to your firewall configuration now. Find the full list of unified outbound IP ranges by region on the Configuring the firewall for Automation Cloud page.
2. Until September 16, 2026: Keep both the current Automation Ops outbound IP ranges (on the Legacy - Configuring the firewall for Automation Cloud page) and the new unified IP ranges allowlisted simultaneously.
3. After September 16, 2026: Continue allowlisting both sets of IP ranges until a subsequent release note confirms the legacy IP ranges can be safely removed.

5 juin 2026

Enregistrement dynamique du client MCP pour les politiques AI Trust Layer [Aperçu]

Vous pouvez désormais activer l'enregistrement dynamique du client MCP comme bascule de fonctionnalité dans une politique AI Trust Layer. Lorsqu’il est activé, AI Trust Layer fournit un enregistrement dynamique du client sécurisé pour les clients MCP, simplifiant leur intégration et la gestion du cycle de vie. Pour de plus amples informations, consultez la section Paramètres des politiques AI Trust Layer.

LLM region routing through the AI Trust Layer policy [Preview]

You can now restrict which geographic regions LLM traffic is routed to using the new Model Regions tab in the AI Trust Layer policy. Set the LLM Region routing toggle to Custom to manually select one or more regions from a list of supported locations worldwide.

Products that use LLMs routed outside the selected regions are blocked. For details, see Settings for AI Trust Layer Policies.

2 juin 2026

Prise en charge de la région Corée du Sud pour Automation Ops

Dans le cadre de notre expansion Cloud en cours, nous apportons la souveraineté complète des données à la région Corée du Sud pour Automation Ops. Cela garantit que vos données sont stockées et répliquées en toute sécurité dans plusieurs zones de disponibilité en Corée du Sud, offrant un contrôle total et une confiance en matière de conformité. Pour plus d’informations, reportez-vous à la page Régions cloud globales .