April 2025 Summaries

AI agents are systems that utilize large language models to plan and execute actions. They will significantly impact the workforce in 2025, affecting productivity and efficiency. Building agentic systems is a complex field with challenges such as developing reasoning tasks-specialized models and managing memory effectively. Memory is crucial for AI agents to learn from past interactions, retain information, and maintain context. AI agents rely on both short-term and long-term memory, with short-term memory working like computer RAM holding onto relevant details for an ongoing task or conversation. Long-term memory works more like a hard drive, storing vast amounts of information to be accessed later. Managing long-term memory is complex due to challenges such as deciding which type of memories to store and retrieve effectively into working memory. Redis stands out as the ideal data platform for agentic memory due to its ability to efficiently store and retrieve memories quickly.

Kong API Gateway and Redis integration is a powerful combination that enhances API management across three main groups of use cases. Kong supports multiple types of Redis deployments for all use cases, including Redis Community Edition, Redis Software, and Redis Cloud. The Kong AI Gateway plugin leverages the existing Kong extensibility model to provide specific AI-based plugins, including semantic caching and semantic routing. These plugins utilize Redis vector similarity search capabilities to implement powerful policies and reduce complexity in deploying AI Gateway capabilities. The combination of Kong and Redis enables developers to build scalable and efficient API management systems with advanced AI capabilities.

The Redis Community and Redis have identified and remediated a security vulnerability, CVE-2025-21605, which allows an unauthenticated client to abuse the output buffer, causing a denial-of-service (DoS) attack. This vulnerability affects all versions of Redis Software and OSS/CE/Stack releases, with fixed releases available in 7.22.0-28 and above for Software, and 7.4.3 and above for OSS/CE. Exposure to this vulnerability requires a publicly exposed Redis endpoint. The community thanks researchers who identified and reported the vulnerabilities through their published process. To protect against this vulnerability, users are advised to follow best practices and upgrade their Redis to the latest release.

The partnership between Unstructured and Redis aims to simplify data ingestion, transformation, and retrieval for organizations building AI workflows. Unstructured's data preprocessing capabilities are combined with Redis' real-time AI capabilities to provide an optimized solution for retrieving augmented generation (RAG) pipelines and other AI-driven apps. This integration simplifies the challenge of preparing diverse data sources, transforming messy data into streamlined formats that can be immediately usable for AI and machine learning tasks. The partnership delivers end-to-end efficiency, scalability, and real-time insights, making it easier for companies to build high-quality machine learning models and apps. The integration is configured through a Redis Cloud destination connector in Unstructured, allowing users to seamlessly integrate their data preprocessing and vector search capabilities.

In the API economy, milliseconds matter. Apigee, a full-featured API management solution from Google Cloud, addresses security, traffic management, and developer adoption challenges. However, as API usage grows, organizations often encounter performance bottlenecks, scalability limitations, high availability requirements, latency issues in global deployments, and evolving API needs. Redis Cloud solves these challenges by offering maximum uptime, peak performance, geographic distribution, scalability, lower TCO, and expert guidance. By deploying Redis Cloud with Apigee, teams can future-proof their API strategy, adapt to changing needs, handle traffic spikes, expand globally, and take advantage of technological advancements while minimizing long-term infrastructure costs. Implementing Redis Cloud is straightforward, requiring three steps: deployment through GCP Marketplace, configuration for caching, rate limiting, and data processing, and monitoring performance improvements through the unified observability dashboard.

Google's Contact Center AI (CCAI) has transformed how businesses engage with their customers, using conversational AI to build intelligent virtual agents. However, even the most powerful platforms face challenges such as latency and complex workflows. Redis Cloud and Arhasi offer targeted solutions to elevate CCAI beyond its core capabilities. Redis Cloud provides a real-time engine that powers CCAI's core, solving performance issues like contextual memory and latency reduction, API and database optimization, and real-time data insights. Arhasi extends CCAI's capabilities by providing turnkey, customizable AI agents with modular templates, enhanced personalization and engagement, intelligent routing, and ecosystem integration. By using Redis Cloud and Arhasi, businesses can deliver faster, smarter experiences while reducing costs and complexity.

Azure Managed Redis is a fully-managed, scalable, in-memory data store that supports both traditional caching and caching for AI apps and workloads. It offers vector data structures and vector search, secondary indexing for full-text search, geospatial queries, numeric data handling, and fast data processing. Azure Managed Redis provides a simplified deployment workflow with a simplified SKU structure based on different performance requirements, allowing customers to choose the right SKU for their workload and network requirements. The service supports password-free authentication via Microsoft Entra ID and enables seamless scaling operations through the portal and API. It integrates seamlessly with key Azure services like Azure SQL, Azure Cosmos DB, and platforms like Azure AI Foundry, enabling optimized "better together" workflows for enhanced performance and efficiency. The service is available in public preview and can be created within minutes starting at 500MB for less than $12 per month.

Redis Cloud has tripled productivity for developers by integrating Redis Insight, a popular developer tool with an intuitive GUI and advanced CLI, seamlessly into its service. This allows users to manage databases, inspect data, and explore data path capabilities within the same context. Redis Insight offers features like list or tree view browsing, full CRUD support, and formatters to make key-value data structures human-readable. The advanced CLI in Workbench provides syntax highlighting, auto-completion, and command browsing, while supporting visualization of command results and time series data. Developers can discover Redis features through interactive tutorials and sign up for Redis Cloud with a free database to get started quickly.

Redis has introduced vector sets, a new data type designed for vector similarity, in its Redis 8 Community Edition. Vector sets are inspired by sorted sets and allow the storage and querying of high-dimensional vector embeddings crucial for various AI and machine learning applications. They offer a simple and intuitive API, reflecting Redis's philosophy of delivering high-performance solutions with minimal complexity. Vector sets can be used to store and retrieve vector embeddings for text descriptions or images, enabling efficient similarity searches. The new data type complements the existing powerful vector search in Redis (Redis Query Engine) and provides a lightweight alternative for specific use-cases. Redis now offers two complementary search capabilities: the Redis Query Engine for comprehensive searching and the vector set for specialized vector similarity search.

Redis has announced two new offerings for AI developers: Redis LangCache, a fully-managed semantic caching service that makes AI apps faster and more accurate, and Vector sets, a native data type that allows developers to easily access and work with vectors. Both of these features are part of a comprehensive real-time data architecture that enables the development of GenAI applications. Additionally, Redis is expanding its GenAI ecosystem with new tools and integrations, including hybrid search, quantization, and support for int8 as an even more memory-efficient vector type. The company is also releasing Redis Agent Memory Server, which provides memory management for AI apps and agents, and Redis Cloud Admin API MCP Server, a natural-language Redis Cloud administrator. Furthermore, Redis Data Integration and Redis Flex are now available on Cloud Pro and Essentials respectively, offering change data capture and fast speeds across both RAM and SSD. Redis Insight is also available in public preview, allowing users to view, update, query, and search the data in their Redis database directly from their browser. Finally, Redis is making it possible to Bring Your Own Cloud and run Redis Cloud in your own Virtual Private Cloud today.

Redis releases for March have introduced features such as preconfigurable database connections using environment variables or a JSON file, enabling centralized and efficient configuration of Redis databases. The company has also improved its data pipeline setup by allowing users to test connectivity to their source database before setting up an RDI data pipeline in Redis Insight. Additionally, Private Service Connect (PSC) connectivity is now supported for Active-Active subscriptions, allowing customers to automate configurations through the Cloud API and Terraform. Furthermore, the Cloud API has been updated to allow extraction of user session logs, with plans to introduce other log categories in the future.