Liulishuo Cuts HarmonyOS Image Load 40% with ImageKnifePro

Challenges in Early HarmonyOS Development

During initial development of the "Liulishuo English" HarmonyOS app, the team faced three major challenges. First, high-concurrency rendering bottlenecks: in scenarios like course list scrolling and word card switching, native components lacked robust memory and disk caching, causing duplicate network requests, wasted bandwidth, and UI stutter. Second, out-of-memory (OOM) risk: long lists and large images without efficient memory reclamation led to crashes, with overall app memory usage under optimization pressure. Third, high cross-platform migration barrier: the team was accustomed to Android's Glide library with its chainable API and lifecycle management, but HarmonyOS lacked a comparable mature framework. Additionally, the app required complex image transformations such as masking, cropping, and blurring; rewriting these with native code would have entailed massive adaptation effort and low reusability.

ImageKnifePro Integration: Architecture and Performance

To address these issues, Liulishuo adopted ImageKnifePro, an open-source image loading library co-built by Huawei and ecosystem partners. ImageKnifePro provides a multi-level caching strategy that enables persistent storage and efficient retrieval of image resources, eliminating redundant downloads. It supports fine-grained LRU memory caching with lifecycle management, releasing resources of destroyed components promptly. Its highly customizable architecture allows custom network stacks, decoders, and interceptor patterns, letting the Liulishuo team reuse existing authentication headers, file caching, and other business logic, fully satisfying architectural decoupling requirements.

Measured Results

After integrating ImageKnifePro, the Liulishuo HarmonyOS app achieved significant gains in both performance and development efficiency:

First-load latency for key images (course banners, word cards) decreased by 40% on average.

Thanks to the powerful caching mechanism, subsequent loads take under 50 ms, delivering near-instant display.

Duplicate download rate dropped by over 30%, and overall app memory usage fell by more than 20%, completely eliminating OOM risk and ensuring a smooth learning experience.

Because ImageKnifePro supports chainable calls and complex image transformations, the amount of migration code from the Android/iOS image libraries was reduced by 30%, greatly simplifying adaptation work and accelerating time-to-market.

Ecosystem Co-Building and Availability

Liulishuo's success validates the core value of open-source components in boosting HarmonyOS development efficiency and user experience. ImageKnifePro, with its mature architecture and strong performance, is becoming the go-to solution for HarmonyOS developers handling complex image scenarios. The library is now published in the OpenHarmony third-party library center; developers can access it at

https://ohpm.openharmony.cn/#/cn/detail/@ohos%2Fimageknifepro

for full source code and documentation. Huawei reports that it has jointly built 90+ scenario-based innovation capabilities with partners, helping 1,200+ partners improve development efficiency and optimize app experience. The company plans to continue its "co-building innovation" approach, releasing more scenario-specific developer solutions across foundational technologies, general capabilities, and vertical industries to further enhance HarmonyOS app innovation and development efficiency.