Android Baseline Profiles: Pre-compiling for Speed

Why it exists

Android's runtime (ART) compiles code as it's needed (Just-In-Time, or JIT), which can cause slowdowns and UI stutter, especially on first launch. Compiling everything Ahead-Of-Time (AOT) makes app installs huge and slow. Baseline Profiles solve this by identifying a small, critical set of code for AOT compilation, giving you the speed of AOT without the size penalty.

The mental model

Think of it as a mise en place for your app. Instead of a chef chopping vegetables for every order (JIT), you give them a prep list for your most popular dishes. Those ingredients are prepped ahead of time (AOT via Baseline Profile), making service instant. Less common orders are still prepped on the fly.

How it works

You define critical user journeys (CUJs) in automated tests, like launching the app and scrolling a list. When you run a special test to generate a profile, the system records every class and method touched during these journeys. This list is saved as a text file (baseline-prof.txt) in your app's source. When a user installs your app from the Play Store, the Android system uses this profile to AOT-compile the listed code, either at install time or during background device idle time.

When to use it

Use Baseline Profiles in any production Android app to improve startup performance and reduce jank. Google reports it can improve time to initial display by up to 40%. It is especially effective for apps with complex startup sequences or performance-sensitive UI, like scrollable lists. It's a standard part of modern Android performance optimization.

When not to use it

There are few reasons not to use them for a production app. The main overhead is setting up the generation process in your CI pipeline. For very simple apps or during early-stage debugging, the performance gain might not be worth the initial setup effort. The optimization is primarily handled by the system for store-based installs, so it's less impactful for apps distributed outside of Google Play.

One canonical example

A social media app defines a critical journey: the user launches the app, lands on their feed, and scrolls down ten items. The developer writes an instrumented test that automates this flow. Generating a Baseline Profile from this test creates a list of all methods involved in app initialization, view inflation, data fetching, JSON parsing, and RecyclerView rendering for those first ten items. This ensures the user's first interaction with the feed is as smooth as possible.