Amazon Elastic Container Registry (Amazon ECR) has increased the maximum number of replication rules per registry from 10 to 25.
Previously, customers with complex multi-region or multi-account architectures were constrained to 10 replication rules per registry, requiring them to consolidate replication configurations to work within that constraint. With this update, customers can define up to 25 replication rules per registry, enabling more precise replication strategies for use cases like distributing images across many regions for low-latency pulls, or replicating to multiple production and staging accounts.
This service limit increase is available in all AWS Regions where Amazon ECR is supported. To learn more, visit the Amazon ECR product page and refer to the Amazon ECR User Guide.
Amazon Elastic Container Registry (Amazon ECR) has increased the maximum number of replication rules per registry from 10 to 25.
Previously, customers with complex multi-region or multi-account architectures were constrained to 10 replication rules per registry, requiring them to consolidate replication configurations to work within that constraint. With this update, customers can define up to 25 replication rules per registry, enabling more precise replication strategies for use cases like distributing images across many regions for low-latency pulls, or replicating to multiple production and staging accounts.
This service limit increase is available in all AWS Regions where Amazon ECR is supported. To learn more, visit the Amazon ECR product page and refer to the Amazon ECR User Guide.
Amazon Bedrock now supports the OpenAI GPT-5.6 models (Sol, Terra, and Luna) on the bedrock-runtime endpoint, with support for the Responses, Converse, and Chat Completions APIs. It also adds support for cross-Region inference, allowing customers to use Global and Geo cross-Region inference to access higher throughput and lower inference costs.
Cross-Region inference automatically routes inference requests across multiple AWS Regions to give you higher throughput, without you needing to manage capacity across multiple Regions. Geo cross region inference routes requests within a predefined geography—including new US Geo (US CRIS) support with this launch—so you can scale while keeping data processed within that geography, while Global cross region inference serve requests from any commercial AWS Region where the model is available, giving you the broadest access to Bedrock capacity and the highest throughput during demand spikes. With Global cross-Region inference you also get lower costs as Global inferencing is priced lower per token for OpenAI models than in-Region and Geo inferencing. This launch also expands API support—you can also use OpenAI GPT models with the Responses API, Chat Completions API, and the Converse API on the bedrock-runtime endpoint. Because these native OpenAI APIs now run on bedrock-runtime, the models work with the same account-level controls you already use for other models on Bedrock: usage appears in Bedrock model invocation logging (deliverable to Amazon S3 or Amazon CloudWatch Logs) and in Amazon CloudWatch metrics covering invocation counts, token counts, latency, throttles, and errors, and it is itemized in AWS Cost Explorer and the AWS Cost and Usage Report so you can attribute spend by model.
Cross-Region inference for OpenAI models is available in all AWS Regions where OpenAI models on Amazon Bedrock are offered. To get started, review the model cards for GPT 5.6 (Sol, Tera and Luna) in the Amazon Bedrock User Guide.
Amazon Bedrock now supports the OpenAI GPT-5.6 models (Sol, Terra, and Luna) on the bedrock-runtime endpoint, with support for the Responses, Converse, and Chat Completions APIs. It also adds support for cross-Region inference, allowing customers to use Global and Geo cross-Region inference to access higher throughput and lower inference costs.
Cross-Region inference automatically routes inference requests across multiple AWS Regions to give you higher throughput, without you needing to manage capacity across multiple Regions. Geo cross region inference routes requests within a predefined geography—including new US Geo (US CRIS) support with this launch—so you can scale while keeping data processed within that geography, while Global cross region inference serve requests from any commercial AWS Region where the model is available, giving you the broadest access to Bedrock capacity and the highest throughput during demand spikes. With Global cross-Region inference you also get lower costs as Global inferencing is priced lower per token for OpenAI models than in-Region and Geo inferencing. This launch also expands API support—you can also use OpenAI GPT models with the Responses API, Chat Completions API, and the Converse API on the bedrock-runtime endpoint. Because these native OpenAI APIs now run on bedrock-runtime, the models work with the same account-level controls you already use for other models on Bedrock: usage appears in Bedrock model invocation logging (deliverable to Amazon S3 or Amazon CloudWatch Logs) and in Amazon CloudWatch metrics covering invocation counts, token counts, latency, throttles, and errors, and it is itemized in AWS Cost Explorer and the AWS Cost and Usage Report so you can attribute spend by model.
Cross-Region inference for OpenAI models is available in all AWS Regions where OpenAI models on Amazon Bedrock are offered. To get started, review the model cards for GPT 5.6 (Sol, Tera and Luna) in the Amazon Bedrock User Guide.
Amazon EC2 Auto Scaling now supports batch instance termination in a single API call. You can now pass up to 100 instance IDs to the TerminateInstanceInAutoScalingGroup API to terminate them as a batch, reducing the number of API calls needed to scale down your Auto Scaling groups.
Batch termination is designed for workloads that need to rapidly scale down, such as AI/ML training jobs, container orchestrators, or event-driven architectures that spin up large fleets temporarily. All instances in a batch are validated atomically before termination begins, and existing Auto Scaling behaviors such as lifecycle hooks and load balancer connection draining are preserved for each instance in the batch.
This feature is available in all AWS Regions at no additional cost.
Amazon EC2 Auto Scaling now supports batch instance termination in a single API call. You can now pass up to 100 instance IDs to the TerminateInstanceInAutoScalingGroup API to terminate them as a batch, reducing the number of API calls needed to scale down your Auto Scaling groups.
Batch termination is designed for workloads that need to rapidly scale down, such as AI/ML training jobs, container orchestrators, or event-driven architectures that spin up large fleets temporarily. All instances in a batch are validated atomically before termination begins, and existing Auto Scaling behaviors such as lifecycle hooks and load balancer connection draining are preserved for each instance in the batch.
This feature is available in all AWS Regions at no additional cost.
To learn more, visit Amazon EC2 Auto Scaling User Guide and Amazon EC2 Auto Scaling API Reference Guide.
Starting today, Amazon Elastic Compute Cloud (Amazon EC2) R8i and R8i-flex instances are available in the Canada West (Calgary) region. These instances are powered by custom Intel Xeon 6 processors, available only on AWS, delivering the highest performance and fastest memory bandwidth among comparable Intel processors in the cloud. The R8i and R8i-flex instances offer up to 15% better price-performance, and 2.5x more memory bandwidth compared to previous generation Intel-based instances. They deliver 20% higher performance than R7i instances, with even higher gains for specific workloads. They are up to 30% faster for PostgreSQL databases, up to 60% faster for NGINX web applications, and up to 40% faster for AI deep learning recommendation models compared to R7i.
R8i-flex, our first memory-optimized Flex instances, are the easiest way to get price performance benefits for a majority of memory-intensive workloads. They offer the most common sizes, from large to 16xlarge, and are a great first choice for applications that don’t fully utilize all compute resources.
R8i instances are a great choice for all memory-intensive workloads, especially for workloads that need the largest instance sizes or continuous high CPU usage. R8i instances offer 13 sizes including 2 bare metal sizes and the new 96xlarge size for the largest applications. R8i instances are SAP-certified and deliver 142,100 aSAPS, the highest among all comparable machines in on-premises and cloud environments, delivering exceptional performance for mission-critical SAP workloads.
To get started, sign in to the AWS Management Console. Customers can purchase these instances via Savings Plans, On-Demand instances, and Spot instances. For more information about the new R8i and R8i-flex instances visit the AWS News blog.
Starting today, Amazon Elastic Compute Cloud (Amazon EC2) R8i and R8i-flex instances are available in the Canada West (Calgary) region. These instances are powered by custom Intel Xeon 6 processors, available only on AWS, delivering the highest performance and fastest memory bandwidth among comparable Intel processors in the cloud. The R8i and R8i-flex instances offer up to 15% better price-performance, and 2.5x more memory bandwidth compared to previous generation Intel-based instances. They deliver 20% higher performance than R7i instances, with even higher gains for specific workloads. They are up to 30% faster for PostgreSQL databases, up to 60% faster for NGINX web applications, and up to 40% faster for AI deep learning recommendation models compared to R7i.
R8i-flex, our first memory-optimized Flex instances, are the easiest way to get price performance benefits for a majority of memory-intensive workloads. They offer the most common sizes, from large to 16xlarge, and are a great first choice for applications that don’t fully utilize all compute resources.
R8i instances are a great choice for all memory-intensive workloads, especially for workloads that need the largest instance sizes or continuous high CPU usage. R8i instances offer 13 sizes including 2 bare metal sizes and the new 96xlarge size for the largest applications. R8i instances are SAP-certified and deliver 142,100 aSAPS, the highest among all comparable machines in on-premises and cloud environments, delivering exceptional performance for mission-critical SAP workloads.
To get started, sign in to the AWS Management Console. Customers can purchase these instances via Savings Plans, On-Demand instances, and Spot instances. For more information about the new R8i and R8i-flex instances visit the AWS News blog.
Anunciamos MAI-Cyber-1-Flash dentro de MDASH, nuestro arnés multiagente para identificar y remediar vulnerabilidades. Juntos ofrecen un rendimiento de clase mundial al 50% del coste de los modelos líderes.
El progreso en IA ha sido sorprendente, al igual que la nueva generación de amenazas cibernéticas que ha comenzado a desatar. Los atacantes ahora poseen capacidades cada vez más poderosas, al explorar una montaña de código cada vez mayor en busca de una sola debilidad que les permita entrar.
A medida que el coste de encontrar un fallo se derrumba, el antiguo modelo de seguridad, donde a veces escaneaban y luego parchaban de manera eventual, ahora queda obsoleto. Si queremos desbloquear los verdaderos beneficios de la IA, primero debemos construir modelos cibernéticos excepcionales que nos ayuden a todos a reforzar el software con el que funciona el mundo.
Esa es la motivación detrás de MAI-Cyber-1-Flash, que ha sido diseñado para encontrar vulnerabilidades desafiantes en bases de código complejas. Ha sido integrado a profundidad en MDASH, perfeccionado por los mejores expertos en ciberseguridad del sector y reforzado en el mayor sector de seguridad del planeta.
Esta experiencia combinada ofrece una protección de seguridad excepcional, que superan a Mythos, Gemini y GPT en CyberGym, el estándar de referencia para evaluar cómo los sistemas razonan sobre grandes bases de código para encontrar vulnerabilidades reales en el código.
Elegir el modelo adecuado para la tarea
La seguridad es una misión siempre activa, y dado el enorme volumen de ataques entrantes, el coste de tokens es ahora la verdadera limitación para los defensores. MAI-Cyber-1-Flash fue diseñado para gestionar de manera eficiente hasta el 90% de todas las tareas, lo que permite a MDASH utilizar los modelos más grandes y costosos de nuestra flota (en este caso GPT-5.4) para el 10% de tareas de una dificultad excepcional, que en verdad las necesitan.
El resultado es que el sistema unificado de MDASH con MAI-Cyber-1-Flash ofrece un 96% en CyberGym (en cualquier puntuación de fallo del benchmark; supera a Mythos en la clasificación de CyberGym).
Esta combinación supone un ahorro del 50% en comparación con nuestra mejor oferta actual en MDASH (GPT 5.4 + 5.4 mini + códex 5.3). Esa es la potencia de un sistema bien ajustado y multimodelo, con acceso a datos históricos de entrenamiento únicos y ricos. Garantiza que ustedes siempre tengan el mejor modelo al mejor precio para cada tarea.
En este nuevo entorno, poder pasar de identificar una nueva vulnerabilidad a abordarla en tiempo real es fundamental. Y aunque la remediación de vulnerabilidades de software por IA es ahora un flujo de trabajo clave de seguridad, hay muchos trabajos que los propios profesionales de seguridad pueden realizar.
Por eso también lanzamos Perception, nuestros sistemas de seguridad agéntica, que proporcionan equipos de agentes para una variedad de flujos de trabajo de seguridad en MDASH, para monitorizar, parchear y cerrar de manera continua nuevos vectores de amenaza. Perception también utilizará pronto MAI-Cyber-1-Flash para muchos más flujos de trabajo de seguridad, más allá del trabajo de vulnerabilidades de software.
Tres cosas importan hoy: Modelo. Datos. Arnés.
Hemos optimizado de manera conjunta nuestros modelos de clase mundial, nuestros datos históricos inigualables y nuestro equipo optimizado por expertos para garantizar que nuestros clientes cuenten con una oferta de seguridad única y potente.
Modelo. MAI-Cyber-1-Flash es un modelo de seguridad compacto y cargado de código derivado de la línea MAI-Thinking-1, que se construyó desde cero, de manera interna, con datos de la más alta calidad. Detalles en nuestro informe técnico.
Datos. Nuestra mayor ventaja. Décadas de construcción de sistemas de seguridad de clase mundial nos ofrecen ahora billones de señales diarias en identidad, terminales, nube y red, y un historial inigualable de exploits y remediaciones reales. Nadie puede fabricar esta historia.
Arnés. MDASH, nuestro arnés multiagente para identificación y remediación de vulnerabilidades, está optimizado por los mejores expertos en seguridad del sector, que han creado 100+ agentes por medio de múltiples modelos líderes para encontrar, validar y remediar vulnerabilidades. El escaneo de código agente es una función crítica en el Centro Operativo de Seguridad y alimenta el Proyecto Perception, nuestro nuevo sistema de seguridad agéntica.
Construido con la seguridad primero
Como MAI-Cyber-1-Flash es el primer modelo cibernético de Microsoft, hemos incorporado confianza en cada capa del sistema, desde el entrenamiento del modelo hasta el despliegue del cliente. El modelo fue desarrollado con una calibración centrada en la seguridad, evaluado de manera rigurosa por el equipo rojo de IA de Microsoft, probado mediante ejercicios adversariales automatizados y dirigidos por expertos, y evaluado de manera independiente por un tercero.
La confianza va más allá del propio modelo. A través de MDASH, los clientes obtienen controles de nivel empresarial, incluidos controles basados en roles, aislamiento de inquilinos, cifrado, auditabilidad y entornos de ejecución en formato sandbox, sin acceso a internet. El resultado es un modelo cibernético que ofrece potentes capacidades a los defensores, para mantener al mismo tiempo la gobernanza, la seguridad y el control que las empresas esperan de Microsoft.
Nuestra máquina de subir colinas
La ciberseguridad no es solo un ámbito rico en datos; es un bucle de aprendizaje por refuerzo en vivo. Cada día, los defensores investigan amenazas, clasifican alertas, cazan adversarios, remedian vulnerabilidades, despliegan protecciones y aprenden del resultado.
Microsoft ve ese bucle de extremo a extremo: vulnerabilidades a través del Centro de Respuesta de Seguridad de Microsoft; ataques y defensas en identidad, terminales, nube, datos, navegador y aplicaciones; más de 100 billones de señales de seguridad cada día; e información operativa de 1,6 millones de clientes. Porque podemos relacionar las acciones con los resultados; lo que era explotable, lo que estaba contenido, lo que estaba bloqueado y lo que en verdad funcionaba; tenemos más que datos.
Nuestro bucle de aprendizaje por refuerzo MAI nos da la base para construir modelos cibernéticos que mejoran de manera continua y se conviertan en defensores expertos en ciberseguridad. Ese será nuestro compromiso con nuestros clientes durante muchos años.
Actualizado a 13 de agosto de 2026
Aclaración sobre las puntuaciones de CyberGym
Any-crash: mide la capacidad del agente para identificar vulnerabilidades que puedan hacer que el código en evaluación se bloquee con una entrada que desencadene cualquier vulnerabilidad existente o de día 0. Nuestra puntuación del 96% es de cualquier accidente
Objetivo (Cualquiera de): mide la capacidad del agente para generar una o más vulnerabilidades candidatas que activen entradas con al menos uno de los candidatos asignando una vulnerabilidad conocida en la suite de pruebas CyberGym. En esta medida, nuestra puntuación es del 90,4%
Envío final: nuevo mecanismo de puntuación introducido en julio. Se basa en el método any-of y pide al agente que elija solo una vulnerabilidad que active la entrada, que se compara con las vulnerabilidades conocidas en la suite de pruebas de CyberGym. Nuestras puntuaciones adoptan un enfoque conservador y filtran los casos límite que el evaluador de Cybergym puede interpretar de manera incorrecta como accidentes válidos. El 86,3% en la clasificación de CyberGym corresponde a la puntuación final de la presentación
AWS CloudShell now includes a built-in visual file editor that you can launch directly from your shell session using a single ‘edit’ command, no setup required. CloudShell provides a browser-based shell environment that uses your existing AWS Management Console credentials, making it a popular choice for developers, DevOps engineers, and cloud administrators managing scripts, infrastructure-as-code, agentic workflows, and AWS Lambda workflows. This new capability extends CloudShell’s utility by bringing a familiar, GUI-based editing experience directly into your console session.
Previously, editing files in CloudShell required using terminal-based editors such as Vim or Emacs, or downloading files locally and re-uploading them after making changes. This added friction to common workflows. The built-in visual editor eliminates this by supporting syntax highlighting, find-and-replace, multi-line selection, copy-paste, and undo-redo in a single browser session. Whether you are updating a deployment script, modifying an agent steering file, editing an AWS CloudFormation template, or fixing an AWS Lambda function, the editor enables a seamless edit-and-run experience.
This feature is available in all AWS Regions where AWS CloudShell is available.
AWS CloudShell now includes a built-in visual file editor that you can launch directly from your shell session using a single ‘edit’ command, no setup required. CloudShell provides a browser-based shell environment that uses your existing AWS Management Console credentials, making it a popular choice for developers, DevOps engineers, and cloud administrators managing scripts, infrastructure-as-code, agentic workflows, and AWS Lambda workflows. This new capability extends CloudShell’s utility by bringing a familiar, GUI-based editing experience directly into your console session.
Previously, editing files in CloudShell required using terminal-based editors such as Vim or Emacs, or downloading files locally and re-uploading them after making changes. This added friction to common workflows. The built-in visual editor eliminates this by supporting syntax highlighting, find-and-replace, multi-line selection, copy-paste, and undo-redo in a single browser session. Whether you are updating a deployment script, modifying an agent steering file, editing an AWS CloudFormation template, or fixing an AWS Lambda function, the editor enables a seamless edit-and-run experience.
This feature is available in all AWS Regions where AWS CloudShell is available.
To learn more about the built-in visual file editor and how to get started, visit the AWS CloudShell documentation.
Amazon Quick announces the general availability of Microsoft 365 extensions for Excel, PowerPoint, Word, and Outlook. These extensions enable Quick to perform tasks directly within users’ M365 environments, using AI to handle complex local tasks such as redlining documents, building financial models, creating presentation-ready decks, and managing Outlook inboxes.
The Excel extension helps with complex spreadsheet analysis, creating pivot tables and charts, and importing and cleaning data. The PowerPoint extension helps you create and refine presentations from Quick data using organization-defined templates. The Word extension generates formatted documents with Word primitives, makes sweeping edits with track changes enabled, and participates as a reviewer in comments. The Outlook extension performs inbox and calendaring tasks such as prioritizing emails, organizing your inbox, scheduling meetings, and drafting replies using your Quick data and entire inbox context.
These extensions transform daily work across teams. Finance teams can build complex models by describing what they need. Sales teams can draft proposals that automatically pull from CRM data. Marketing teams can create branded presentations without manual formatting. Legal teams can streamline contract reviews. Operations teams can manage email workflows and schedule meetings intelligently, and IT teams can automate routine data analysis that previously required manual effort.
Amazon Quick Microsoft 365 extensions are available in US East (N. Virginia), US West (Oregon), Asia Pacific (Sydney), Europe (Ireland), Asia Pacific (Tokyo), and Europe (Frankfurt). To learn more, see Amazon Quick for Microsoft 365: Agentic AI where you work, and download extensions on the Quick download page.
Amazon Quick announces the general availability of Microsoft 365 extensions for Excel, PowerPoint, Word, and Outlook. These extensions enable Quick to perform tasks directly within users’ M365 environments, using AI to handle complex local tasks such as redlining documents, building financial models, creating presentation-ready decks, and managing Outlook inboxes.
The Excel extension helps with complex spreadsheet analysis, creating pivot tables and charts, and importing and cleaning data. The PowerPoint extension helps you create and refine presentations from Quick data using organization-defined templates. The Word extension generates formatted documents with Word primitives, makes sweeping edits with track changes enabled, and participates as a reviewer in comments. The Outlook extension performs inbox and calendaring tasks such as prioritizing emails, organizing your inbox, scheduling meetings, and drafting replies using your Quick data and entire inbox context.
These extensions transform daily work across teams. Finance teams can build complex models by describing what they need. Sales teams can draft proposals that automatically pull from CRM data. Marketing teams can create branded presentations without manual formatting. Legal teams can streamline contract reviews. Operations teams can manage email workflows and schedule meetings intelligently, and IT teams can automate routine data analysis that previously required manual effort.
Amazon Quick Microsoft 365 extensions are available in US East (N. Virginia), US West (Oregon), Asia Pacific (Sydney), Europe (Ireland), Asia Pacific (Tokyo), and Europe (Frankfurt). To learn more, see Amazon Quick for Microsoft 365: Agentic AI where you work, and download extensions on the Quick download page.
Amazon OpenSearch Service now extends automatic semantic enrichment to VPC-enabled domains, allowing customers with private network configurations to leverage AI-powered semantic search without exposing their domains to the public internet.
Automatic semantic enrichment transforms traditional keyword-only search into context-aware retrieval by understanding the meaning behind queries. For example, a search for «lightweight laptop for travel» returns results about «ultrabooks» and «portable notebooks under 3 lbs» even when these exact terms aren’t in the query. The feature handles all semantic processing automatically, eliminating the need to self-manage machine learning models and integration overhead. Previously, automatic semantic enrichment was available only on domains that are not VPC-enabled. This capability is now supported within VPCs (Virtual Private Clouds), enabling customers with stricter network security requirements to improve search relevance while maintaining their existing security posture. No changes to existing VPC configurations are required. To learn more about automatic semantic enrichment, see our documentation.
Automatic semantic enrichment for VPC domains is available across 11 Regions globally: US East (N. Virginia, Ohio), US West (Oregon), Asia Pacific (Mumbai, Singapore, Sydney, Tokyo), and Europe (Frankfurt, Ireland, Spain, Stockholm). To get started, on your VPC-enabled OpenSearch Service domain running OpenSearch version 2.19 or later, create an index with automatic semantic enrichment fields configured. Note that you may need to update your domain to the latest service software version, see updating service software for more information.
Amazon OpenSearch Service now extends automatic semantic enrichment to VPC-enabled domains, allowing customers with private network configurations to leverage AI-powered semantic search without exposing their domains to the public internet.
Automatic semantic enrichment transforms traditional keyword-only search into context-aware retrieval by understanding the meaning behind queries. For example, a search for «lightweight laptop for travel» returns results about «ultrabooks» and «portable notebooks under 3 lbs» even when these exact terms aren’t in the query. The feature handles all semantic processing automatically, eliminating the need to self-manage machine learning models and integration overhead. Previously, automatic semantic enrichment was available only on domains that are not VPC-enabled. This capability is now supported within VPCs (Virtual Private Clouds), enabling customers with stricter network security requirements to improve search relevance while maintaining their existing security posture. No changes to existing VPC configurations are required. To learn more about automatic semantic enrichment, see our documentation.
Automatic semantic enrichment for VPC domains is available across 11 Regions globally: US East (N. Virginia, Ohio), US West (Oregon), Asia Pacific (Mumbai, Singapore, Sydney, Tokyo), and Europe (Frankfurt, Ireland, Spain, Stockholm). To get started, on your VPC-enabled OpenSearch Service domain running OpenSearch version 2.19 or later, create an index with automatic semantic enrichment fields configured. Note that you may need to update your domain to the latest service software version, see updating service software for more information.
Amazon Quick now enables administrators to set per-user limits on index storage and agent hours, giving them direct control over subscription costs. With limits management, administrators can create limit profiles that cap per-user consumption, helping prevent unexpected overage charges and ensuring subscription entitlements are used efficiently across their organization.
For example, an organization deploying Quick enterprise-wide to thousands of users can set account-level limit profiles to establish cost-predictable baselines, then assign higher limits to specific roles that require more agent hours. Administrators can create and assign limit profiles at the user, role, or account level, with a priority hierarchy that ensures the right users get the right capacity. When a user reaches their limit, new consumption is blocked while existing content is preserved.
This feature is available for Professional and Enterprise plans, in all AWS Regions where Amazon Quick agentic capabilities are supported. For more information, see the AWS Region table.
Learn more about Amazon Quick by visiting the Quick website. To learn more about limit profiles, see Limits management in the Amazon Quick User Guide.
Amazon Quick now enables administrators to set per-user limits on index storage and agent hours, giving them direct control over subscription costs. With limits management, administrators can create limit profiles that cap per-user consumption, helping prevent unexpected overage charges and ensuring subscription entitlements are used efficiently across their organization.
For example, an organization deploying Quick enterprise-wide to thousands of users can set account-level limit profiles to establish cost-predictable baselines, then assign higher limits to specific roles that require more agent hours. Administrators can create and assign limit profiles at the user, role, or account level, with a priority hierarchy that ensures the right users get the right capacity. When a user reaches their limit, new consumption is blocked while existing content is preserved.
This feature is available for Professional and Enterprise plans, in all AWS Regions where Amazon Quick agentic capabilities are supported. For more information, see the AWS Region table.
Learn more about Amazon Quick by visiting the Quick website. To learn more about limit profiles, see Limits management in the Amazon Quick User Guide.
Amazon Quick now integrates with Microsoft Purview to enforce data loss prevention (DLP) policies across your Quick environment. Organizations need to ensure that sensitive files aren’t shared outside approved channels. With this integration, IT administrators and security teams can apply their existing Purview sensitivity labels to automatically control how files are handled in Quick capabilities such as chat, spaces, and knowledge bases.
Administrators can configure enforcement actions (block, warn, or allow) for each sensitivity label, giving organizations granular control over sensitive file sharing across Quick. For example, a financial services company can block files labeled «Highly Confidential» from being uploaded to shared spaces while allowing «Internal» files with a warning notification. With this integration, customers can extend their existing Microsoft Purview governance policies into Quick without additional tools.
Amazon Quick now integrates with Microsoft Purview to enforce data loss prevention (DLP) policies across your Quick environment. Organizations need to ensure that sensitive files aren’t shared outside approved channels. With this integration, IT administrators and security teams can apply their existing Purview sensitivity labels to automatically control how files are handled in Quick capabilities such as chat, spaces, and knowledge bases.
Administrators can configure enforcement actions (block, warn, or allow) for each sensitivity label, giving organizations granular control over sensitive file sharing across Quick. For example, a financial services company can block files labeled «Highly Confidential» from being uploaded to shared spaces while allowing «Internal» files with a warning notification. With this integration, customers can extend their existing Microsoft Purview governance policies into Quick without additional tools.
This feature is available in all AWS Regions where Amazon Quick agentic capabilities are supported. For more information, see the AWS Region table. To learn more, see Data loss prevention in the Amazon Quick User Guide.
