The first 100% Local Android App Security Analyzer Audit any Android
.apkfile for dangerous permissions, suspicious combinations, and code-level behaviours without ever uploading it. The analysis runs entirely within your browser using a high-performance, zero-knowledge pipeline.
Appranium performs true binary analysis entirely in your browser. It is not a wrapper for a backend API.
- File Ingestion: When an APK is dragged into the app, the HTML5 File API reads it directly into the local memory of the browser instance. No external network request is established.
- In-Memory Extraction: Utilizing
jszip, the engine decompresses the APK archive to isolate critical components, specificallyAndroidManifest.xmlandclasses.dex. - Binary AXML Decompilation: Android apps compile XML into binary formats (AXML). Appranium features a custom JavaScript
axmlParserthat reconstructs the binary string pool and XML tree structure to extract declared permissions, application metadata, and manifest configurations. - True-Binary DEX Byte-Code Parsing: The system performs structural binary parsing on the Dalvik Executable (
classes.dex). Utilizing a customDataViewengine, it deserializes the DEX header, decodes ULEB128 strings, and maps exactClassName->methodNameinvocations. It actively flags dynamic code loading (DexClassLoader), reflection abuse (Method.invoke), shell commands (Runtime.exec), and deep hardware invocations (e.g.AudioRecord). - Contextual Risk Intelligence: Extracted datasets are fed into a weighted scoring matrix. The engine correlates declared permissions against DEX findings and flags suspicious combinations (e.g., recording audio alongside hidden network access) to produce a precise, transparent security grade.
- Parallax Interface: A modern, editorial-grade UI with smooth scroll dynamics and glassmorphism.
- Final Result Animation: A high-impact cinematic reveal of the underlying application's security grade.
- 3D Interactive Elements: Highly polished WebGL and CSS3D elements for a premium feel.
- DEX Byte-Code Insights: Automatic scanning for runtime permission requests and sensitive data sinks.
- Binary AXML Parsing: Professional-grade extraction of manifest data directly from the compiled APK container.
- Simulation Layer: Predictive modeling that flags potential attack vectors based on observed structural patterns.
- Weighted Intelligence: Dynamic calculation weighting individual permissions, combinations, and code-level strings.
- Customizable Tuning: A dedicated configuration modal allowing security professionals to adjust risk sensitivity multipliers.
- Sticky Intelligence Panel: A persistent HUD tracking the running risk score and top threat drivers as you traverse the report.
- Suspicious Combinations: Specific threat intelligence modeling for dangerous pairings (e.g., Camera + Network) often used by APTs or spyware.
- Domain-Based Profile: Visual categorization mapping the app's exposure across Location, Identity, Storage, and Hardware domains.
- Educational Context: "Why it matters" deep-dives for every flagged permission, translating raw data into actionable security intelligence.
- Node.js v18+
- npm v9+
git clone https://github.com/your-username/appranium.git
cd appranium
npm install
npm run devnpm run build
# The 'dist/' directory is ready for static web hosting (Vercel, Netlify, Github Pages)src/
├── components/
│ ├── ParallaxHero/ # Glassmorphic landing hero
│ ├── FinalResultAnimation/ # Cinematic result reveal
│ ├── StickyRiskPanel/ # Persistent risk HUD
│ ├── CodeLevelInsights/ # DEX byte-code analysis view
│ ├── SimulatedRisks/ # Privacy threat simulations
│ ├── RiskScore/ # Weighted intelligence engine UI
│ ├── RiskCharts/ # Recharts visualizations
│ ├── SuspiciousCombos/ # Pattern-match alert system
│ └── common/ # Premium atomic UI primitives
├── hooks/
│ └── useAnalysis.js # Core orchestration & state management
└── utils/
├── axmlParser.js # Binary manifest deconstruction
├── comboDetector.js # Threat pattern matching
├── fileHandler.js # APK structure handling (JSZip)
├── manifestParser.js # Analysis pipeline orchestrator
└── riskScorer.js # Weighted risk logic
Appranium is built on the principle of Zero-Knowledge Analysis.
- Local Execution: All analysis algorithms happen in your browser's V8 Javascript engine.
- No Data Transit: Your
.apkfiles never leave your machine; they don't touch any API. - In-Memory Processing: Decompiled data is held in-RAM and forcefully cleared on session reset.
- Offline Capable: Once the app dashboard loads, the core engine functions 100% offline.
MIT © Appranium Contributors