I will present a technique to speed up subsequent Clang Static Analyzer (CSA) runs on mostly unchanged code. CSA takes a lot more time to complete than simply compiling the source code. This imposes challenges for quick developer feedback, for pull requests, or even within your IDE. For incremental and localized changes, we could reuse the bulk of the previous analysis and only re-analyze the changed parts that actually matter. In this talk, besides presenting this novel technique for incremental analysis, I will summarize how CSA currently selects and analyzes functions and elaborate on how this technique could fit into the current architecture.