San Francisco, November 6, 2026 / PRNewswire / -- Redis Today, the next phase of Redis Iris has been released, which is its real-time context engine for AI agents. Redis Iris now offers six customizable features that can be used individually or in combination, designed to provide agents with the memory, context, and messaging capabilities necessary for reliable operation in real environments: Redis Intent Router, Redis Agent Playbooks, Redis Agent Memory, Redis Context Gateway, Redis Lens, and Redis Agent Mesh.

The company stated that the AI model and its related infrastructure are now powerful enough for enterprise use, but so far only a few teams have achieved repeatable success in production environments with agents. Redis believes that agents often lack the ability to obtain the correct context at the right time and to have appropriate memory. Agents that can create value typically do not operate efficiently: they use overly long prompts, repetitive LLM calls, and fragile custom integrations, and the costs associated with these integrations in terms of token and time often exceed the benefits they bring.

"Proxies require the same foundation as applications and microservices: up-to-date application data, runtime state, and communication between services," said Rowan Trollope, CEO of Redis. "In addition, they also need new capabilities such as memory, context management, and routing. As a leading runtime database for applications, Redis is already on the path to handling time-sensitive tasks. Developers should not need five different infrastructure products just to make proxies work. Iris integrates these fundamental components into one place, allowing developers to build proxies that are ready for use in a production environment."

Six functions, one runtime environment

Iris has provided production-level features for context, memory, and messaging from the very first day of release, and at the same time, the team can also customize each feature according to their own proxies. This release includes:

• Intent Router ( GA ) recognizes the meaning of a request, rather than just its literal expression, and routes it to the correct destination, such as a saved FAQ answer, a proxy skill, a human operator, or LLM. This feature also supports routing to proxies or models, as well as "fuzzy" if statements, allowing decisions to be no longer limited to true or false.
• Agent Playbooks ( GA ) allows teams to define which answers should remain valid for a long time and can be reused, and which answers require the latest data and the re-execution of known processes. Each execution makes playbook more complete and accurate, thereby reducing the time taken to solve problems and lowering token consumption.
• Agent Memory ( GA ) generates memories from user operations, documents, and other semi-structured data through MemoryETL, and organizes them in a hierarchical manner for more precise control. This feature also provides capabilities for viewing, auditing, and proving the source of information, and allows for targeted forgetting of content that should not be retained, such as personal data.
• Context Gateway ( GA ) provides a centrally managed entry point for proxies to access connected context sources. Teams can integrate their own MCP servers to cache tools that can solve specific problems, thereby reducing token consumption and creating reusable results to maintain consistency across different sessions.
• Lens ( Beta ) makes proxy behavior traceable. Each result can be traced back to the skills used, the selected routes or FAQ employed, as well as the memory on which the proxy's learning is based and the initial session. This allows for the generation of auditable evidence and provides insights into task success rates, the token consumption, and delays for each task.
• Agent Mesh ( Beta ) enables agents to communicate through brief streams. If the connection is interrupted, data will be automatically buffered and the connection will be re-established. Durable SSE Streams ensures responsive transmission even during interruptions. A2A Comms provides a reliable channel for agents to delegate and hand over tasks.

The company states that these functions collectively help agents remember the correct information, answer correctly, and maintain traceability in each operation, while producing more results with fewer token. Since Iris is built on the Redis technology that the team is already familiar with, enterprises can reduce the amount of custom integration work that is usually required, which is otherwise used to connect information retrieval, memory, routing, and message delivery.

Designed for real-world challenges

Proxies in a production environment must be capable of handling unpredictable human input, multiple channels, and handovers between proxies, all of which can become points of failure. Iris is based on four principles: context should be navigable rather than isolated in silos; memory should be adaptive and selective; context should be used efficiently to control costs; and as proxies learn from past interactions, the context should continue to add value over time.

Since Redis Iris was launched in May 2026, customers have been using this solution in production environments.

Paradigm Chief Technology Officer June Lee said: "Everything you see in Paradigm is stored in Redis. Redis is the memory and orchestration layer that allows our agents to run in real-time, on a scale that was previously impossible to achieve."

Availability and more information

Redis Iris is now available. Intent Router, Agent Playbooks, Agent Memory, and Context Gateway have been officially launched; Lens and Agent Mesh are in the Beta phase. For more information, please visit redis.io / iris.

Redis will also conduct the first live demonstration of the new version together with the development team: the North American and South American sessions are on November 12th, and the EMEA session is on November 17th. The registration addresses are redis.io / resources / videos / redis-iris-full-spectrum.

About Redis

Redis claims to be the fastest data platform in the world. Since its inception as an open-source project in 2011, Redis has become the most frequently mentioned brand in caching solutions and has helped over 10,000 customers build, scale, and deliver applications that support the operation of the world. With its cloud and local databases for caching, generative AI, and other purposes, Redis helps digital enterprises set new standards in application speed.

Redis has offices in San Francisco, Austin, London, Tel Aviv, and Singapore, and is internationally recognized as a leading provider for rapidly building high-performance applications. For more information, please visit redis.io.