pyjadx

Python bindings for JADX via JPype. Decompile Android APK and DEX files to Java source from Python.

from pyjadx import Jadx

with Jadx("app.apk") as jadx:
    for cls in jadx.classes:
        print(cls.full_name)
        print(cls.code)

Why pyjadx?

JADX is the leading open-source Android decompiler — it produces clean Java/Kotlin source from Dalvik bytecode, preserving class hierarchies, generics, exception handling, and annotations. pyjadx brings this into Python without subprocess overhead: JADX runs in-process via JPype.

Install

pip install pyjadx

Requirements: JDK 11+ (JADX JARs are downloaded automatically on first use).

Quick start

from pyjadx import Jadx

with Jadx("app.apk") as jadx:
    # All top-level classes (inner classes via cls.inner_classes)
    for cls in jadx.classes:
        print(cls.full_name, len(cls.methods), "methods")

    # Search by fully qualified name
    main = jadx.search_class("com.example.MainActivity")
    if main:
        print(main.code)

    # Method-level decompilation
    for method in main.methods:
        print(f"{method.name}: {method.code}")

    # Cross-references
    for method in main.methods:
        print(f"{method.name} called by: {[c.name for c in method.callers]}")

    # Package navigation
    for pkg in jadx.packages:
        print(pkg.name, len(pkg.classes), "classes")

Supported inputs

Format	Example
APK	Jadx("app.apk")
DEX	Jadx("classes.dex")
Multi-DEX APK	Handled transparently — all classes merged

API reference

Jadx(path, *, threads=None, mode="auto", code_cache=False)

Main entry point. Use as a context manager or call load()/close() manually.

Parameter	Description
path	Path to APK or DEX file
threads	JADX thread pool size (None = CPU count)
mode	"auto", "restructure", "simple", or "fallback"
code_cache	False = in-memory cache (default), True = disk-backed

Properties:

Property	Type	Description
classes	list[JavaClass]	Top-level classes across all DEX files
packages	list[JavaPackage]	Package-level navigation
class_count	int	Number of top-level classes
java	JObject	Escape hatch to underlying jadx.api.JadxDecompiler

Methods:

Method	Description
load()	Load and resolve classes (called by __enter__)
close()	Release resources (called by __exit__)
decompile_all()	Eagerly decompile all classes
search_class(name)	Find class by fully qualified name

JavaClass

Property	Type	Description
name	str	Short name ("MainActivity")
full_name	str	Fully qualified ("com.example.MainActivity")
package	str	Package name
code	str	Decompiled Java source (lazy, cached)
smali	str	Smali disassembly (lazy, cached)
methods	list[JavaMethod]	Methods in this class
fields	list[JavaField]	Fields in this class
inner_classes	list[JavaClass]	Inner/anonymous classes
declaring_class	JavaClass | None	Parent class (if inner)
top_class	JavaClass	Outermost enclosing class
dependencies	list[JavaClass]	Classes this class depends on
is_inner	bool	Whether this is an inner class
java	JObject	Escape hatch to jadx.api.JavaClass

JavaMethod

Property	Type	Description
name	str	Method name
full_name	str	Fully qualified with class
code	str	Decompiled source (lazy, cached)
callers	list[JavaMethod]	Methods that call this one (lazy, cached)
callees	list[JavaMethod]	Methods this one calls (lazy, cached)
override_related	list[JavaMethod]	Bidirectional override chain
declaring_class	JavaClass	Class this method belongs to
java	JObject	Escape hatch to jadx.api.JavaMethod

JavaField

Property	Type	Description
name	str	Field name
full_name	str	Fully qualified with class
type	str	Field type as string
callers	list[JavaMethod]	Methods referencing this field (lazy, cached)
declaring_class	JavaClass	Class this field belongs to
java	JObject	Escape hatch to jadx.api.JavaField

JavaPackage

Property	Type	Description
name	str	Full package name
classes	list[JavaClass]	Classes in this package
sub_packages	list[JavaPackage]	Sub-packages
java	JObject	Escape hatch to jadx.api.JavaPackage

Escape hatch

Every wrapper exposes its underlying JADX Java object via .java. This lets you call any JADX API we haven't wrapped yet:

with Jadx("app.apk") as jadx:
    cls = jadx.search_class("com.example.Foo")
    # Call JADX Java API directly
    code_info = cls.java.getCodeInfo()

JVM coexistence

pyjadx coexists with other JPype consumers (e.g., PyGhidra) in the same process. If a JVM is already running, pyjadx.start() adds JADX JARs to the existing classpath. If not, it starts one.

import pyghidra
pyghidra.start()  # starts JVM for Ghidra

import pyjadx
pyjadx.start()    # adds JADX JARs to existing JVM — no conflict

Configuration

Custom JADX installation:

# Via environment variable
# export JADX_HOME=/opt/jadx

# Or at runtime
import pyjadx
pyjadx.start(jadx_home="/opt/jadx")

JAR caching: JADX JARs are auto-downloaded from Maven Central on first use and cached at ~/.pyjadx/jars/<version>/.

License

MIT