5 big updates to the Axoniq platform
Axoniq Framework adds Workflows. Axon Server 26.0.0 ships more stable under load. Axoniq Insights now queryable on all Team plans. One release cycle. Every layer of the stack moves forward.
TL;DR:
Axoniq Framework 5.1 is here. Business processes that used to live across scattered event handlers and state machines now live in a single, readable function. Workflows are the first capability exclusive to Axoniq Framework, and they ship today.
Feature parity with AF4 is complete. Entity-level snapshotting, Sequenced Dead-Letter Queue, reliable replay, and full support for Spring Boot 4 and Java 21 are all included in this release.
Axon Framework 4 reaches end of life June 30, 2026. If you're still on 4.x, it is time to move.
Axoniq Insights is now available on all Team tier plans and above. Your entire event history is queryable in natural language, in seconds, without writing projections or routing questions through engineering.
Axon Server 26.0.0 ships more stable and more predictable at production load: better gRPC reliability, lower CPU overhead, and automated cluster consistency detection.
Axoniq Framework 5.1 is available today
Workflows are here!
If you've been running process orchestration through Sagas — scattered event handlers, state machines spread across files, logic you have to mentally reassemble every time something breaks in production — Workflows replace all of that with a single, readable, top-down function. One place in the code to see a business process from start to finish. The flow is explicit. The state is traceable. The operational opacity that makes Sagas painful to maintain is gone.
This is where the framework goes from here. Workflows are the first major capability exclusive to Axoniq Framework, and they won't be the last.
Full parity with everything you relied on in 4.x — and then some
Axoniq Framework 5.1 also closes the feature gap that held a lot of teams back from moving to Framework 5. Snapshotting now works at the entity level, not just the aggregate. If your domain model is entity-heavy, you've felt the cost of that gap. The Sequenced Dead-Letter Queue is back, preserving ordering within a sequence so failed messages fail and recover together rather than producing inconsistent state. Event Processor reset and replay is fully reliable. Last-minute message payload conversion handles serialization mismatches transparently at the handler boundary. Replay Status Change Handlers give you clean hooks when a replay starts and concludes — no more inferring replay state inside your handlers.
Rounding it out: multi-context event consumption via multi-streamable-event-source, @Namespace for package-level message namespace declarations, and full support for Jackson 2 and 3, Spring Boot 3 and 4, and Java 21.
What is Axoniq Framework?
Axoniq Framework is the commercial evolution of the Axon framework, built on top of Axon Framework 5, which remains Apache 2.0 and free. Read our CTO’s post here about the evolution to Axoniq Framework.
Axon Framework 4 reaches end of life on June 30 2026. If you're still on 4.x, the migration guide is a good place to start thinking through your path. If you need some extra time, contact us to talk about LTS.
Active Intelligence: Axon Server 2026.0.0 + Insights is here
Our engineering teams have been busy in the first quarter of the year and we’ve got a lot released today, so we've broken our content into a few different sections. This article is specifically about our Insights and Server release, but if you're looking for the framework updates scroll all the way to the bottom.
In October we launched an analytics node that made your entire event history queryable in natural language, across billions of events. The response was clear: this needed to be in every deployment. So we made it one. Starting today, Axoniq Insights is available to all customers on our Team tier and above.
Query Your Entire Event History Without Writing a Single Projection
Every question about what your system did, such as why an order stalled, where a process failed, or what happened during an incident, used to route through engineering. Someone scoped it, built a projection, maintained a schema. The data was always there. Getting to it required a sprint.
One of our customers maintained a manually built traditional database view feeding into their analytics tool to run queries against their event data. When new questions came up, they wrote new views. What used to take them weeks to compile now takes seconds. No custom pipelines. No intermediate projections. Nothing to maintain.
Operations, product, support, and anyone who needs answers gets them directly, without engineering involvement. Your team ships instead of answering questions.
AI readiness
When your organization deploys AI agents into production systems, the question you'll face isn't whether the model is capable enough. It's whether you can explain what it did and why.
Most agent frameworks produce technical execution logs. Metadata about infrastructure operations, not business decisions. When something goes wrong, there's no authoritative trail. When a stakeholder asks why the system made a decision, you reconstruct an answer.
Your Axon Server event store already contains the full causal history of every business decision your system has made; in sequence, immutable, and traceable. Axoniq Insights makes that record accessible at machine speed. Every agent recommendation traces back to a specific business event. Every decision is auditable. Agents that fail resume from where they stopped, not from the beginning.
The infrastructure your team already runs is the infrastructure that makes AI safe to put into production. You're not building that capability later. It's already there.
Event store built for production load
Axon Server 2026.0.0 ships a focused set of production improvements — gRPC reliability under load, reduced CPU overhead, automated cluster consistency detection. The system is more stable and more predictable at the throughput levels your team depends on.
From Individual to Enterprise: Axon Server Pricing Tiers Explained
Each tier is purpose-built for the team size and workload it serves — from Individual (free, always) to Enterprise (uncapped throughput, full data sovereignty, no connection to the Axoniq Platform required). When your team spins up an environment, it's already configured for your scale.
Also shipping today: Axoniq Framework 5.1 — AF5 feature parity complete, AF4 EOL June 30, and the introduction of Axoniq Workflows. [Read the announcement →]
Your history has always been there. Now your whole organization can act on it.
[Get started with Insights →] [Read the Axon Server release notes →]


