Migrate react-native-compressor to Nitro Modules for better performance and smaller binary size #381
Description
Title
Migrate react-native-compressor to Nitro Modules for better performance and smaller binary size
Description
Hi 👋
First of all, thank you for maintaining react-native-compressor. It’s a very useful library and widely used in production apps.
I’d like to propose migrating this library to Nitro Modules, similar to recent high-performance React Native libraries such as react-native-nitro-fs and react-native-fetch-nitro.
Motivation
1. Performance improvements
Nitro Modules significantly reduce JS ↔ Native overhead by relying directly on JSI instead of the classic NativeModule bridge.
This would be especially beneficial for CPU-intensive tasks such as:
- Image compression
- Video compression
- Batch file compression
Lower overhead would result in faster execution and more stable performance, particularly on lower-end Android devices.
2. Smaller Android binary (AAB/APK) size
Nitro Modules allow:
- More granular native code inclusion
- Better dead-code elimination
- Reduced reliance on large shared native libraries
This is becoming increasingly important due to stricter Google Play size and native dependency checks.
3. Better alignment with React Native New Architecture
- Nitro Modules fit naturally with JSI, TurboModules, and Fabric
- More future-proof than the legacy NativeModule approach
- Easier to optimize and maintain long-term
4. Improved developer experience
- Cleaner native bindings
- More predictable performance characteristics
- Easier reasoning about threading and memory usage during compression
Current pain points (user perspective)
- Compression can feel slow on some devices
- Native overhead is noticeable when compressing multiple files sequentially
- App size increases when combined with other native-heavy libraries
Proposed approach
- Introduce a Nitro-based implementation (possibly experimental at first)
- Keep the existing implementation for backward compatibility
- Gradually migrate core compression logic to Nitro Modules
Willingness to help
I’d be happy to:
- Test an experimental Nitro implementation
- Share benchmarks (before / after)
- Provide feedback from real-world production usage
Thanks again for your work on this library 🙏
Looking forward to your thoughts.