Materialize

Materialize computes and then continually maintains the results of your SQL, as the underlying data change. One consequence is that it needs to keep quite a lot of data live and ready to go at a moment's notice. Naively, this translates to quite a lot of

In July we continued to invest in running context-engineering workloads reliably at scale, and in making it easier to build on Materialize with coding agents. Here are some of the highlights; for the full list of changes, see the release notes.Sources and

The teams getting their AI spend under control are doing it through context engineering, not usage caps. The key to their success is where that engineering happens: in a live context graph.How agents spend tokensThe cost equation for an AI agent task is s

Materialize is an agent-native platform, and that means different things for different agents. A coding agent operates the platform itself: it models data, deploys changes, and works out why something broke, without a human driving each step. An applicati

Engineers building context pipelines for agents all face the same conundrum: every system must decide where the necessary data transformation work gets done. Whether at read time, at write time, or somewhere in between, a cost must be paid in the form of

"I didn't have time to write a short letter, so I wrote a long one instead." - Mark TwainAs we help engineering teams bring more and more context pipelines into production, it's gotten me thinking about a writing principle I've learned and relearned over

Since our last update, we've shipped new MCP servers for agents and developers, a new CLI for declarative deployments, and Single Sign-On for Self-Managed. Here are some of the highlights; for the full list of changes, see the release notes . MCP Server f

Search maintenance used to be a UX problem. For agents, it is a correctness problem.Large language models reason in language. So the natural way for an agent to ask the world for information is in words: describe the thing it needs, then search for it.Key

We'll write transaction commit logic in a SQL view, which can enable higher throughput than performing transaction commits in the control plane. Incremental view maintenance (IVM) makes the critical path fast enough that the determination of which transac