San Francisco – November 04, 2026 -- Redis has expanded Redis Iris, its real-time context engine for AI agents, to six customizable capabilities, four of them generally available and two in beta. The company targets the inefficiencies that keep agents out of production: over-stuffed prompts, redundant LLM calls and brittle custom integrations that cost more in tokens and time than the value they return.
Four of six Iris capabilities are generally available today
Intent Router matches the meaning of a request rather than its wording and sends it to a remembered FAQ, an agent skill, a human or an LLM. It also supports agent and model routing, guardrails and "fuzzy" if-statements, so decisions are not limited to true or false.
Agent Playbooks lets teams define which answers are stable and reusable, and which need current data and a fresh run of a known procedure. Redis says playbooks become more thorough and accurate the more they run, reducing resolution times and token spend.
Agent Memory builds memories from user actions, documents and other semi-structured data through MemoryETL. It organizes them in hierarchies and provides inspection, audit and provenance. It can also forget data it should not keep, such as personally identifiable information.
Context Gateway gives agents a single, governed entry point to connected context sources. Teams can bring their own MCP servers and cache which tools solve a given problem to reduce token use. They can also build reusable results for consistency across sessions.
Lens and Agent Mesh enter beta with tracing and messaging resilience
Lens traces every agent outcome back to the skill, route or FAQ that served it, the memory it learned from and the source session. It also reports task success, tokens per task and latency, giving auditable evidence of agent behavior.
Agent Mesh lets agents communicate over ephemeral streams and automatically buffers and reconnects when a connection drops. Durable SSE Streams keep responses flowing through interruptions, and A2A Comms provide a channel for delegation and handoffs between agents.
Redis pitches a single stack to replace five infrastructure products
CEO Rowan Trollope said developers should not need five infrastructure products to make agents work. He argued that Redis already sits in the hot path as a runtime database for apps and that Iris brings the new primitives of memory, context management and routing into one place.
Because Iris runs on Redis technology that teams already use, organizations can cut the custom integration work that typically connects retrieval, memory, routing and messaging, the company said. Iris is built on four principles: context should be navigable rather than siloed, memory should be adaptive and selective, context should be efficient to keep costs in check, and context should compound as agents learn.
Customers have run Iris in production since its May 2026 launch
Paradigm is one of those customers. Its CTO, June Lee, said: "Everything you see in Paradigm is stored in Redis." She described it as the memory and orchestration layer that lets the company's agents work in real time.
Redis will host a live first look at the release with its engineering team on Nov. 12 for the Americas and Nov. 17 for EMEA.