AndroidFridaManager

AndroidFridaManager is a Python API designed to simplify the installation and management of Frida on Android devices. It provides an easy-to-use interface for installing and running the latest Frida server, as well as the flexibility to install specific versions as needed.

Key features:

• Frida Server Management: Seamlessly install and run the latest Frida server on your Android device, or choose to install a specific version as required. By default the installed version matches your host frida Python package so that client and server stay in sync.
• Job Management: Execute Frida scripts as independent jobs, managed by the JobManager(). This feature allows for concurrent execution of multiple Frida scripts, with each job running in its own thread.
• Bundle API: Load a declarative set of hook-sets under one opaque bundle id. On a paused spawn all scripts are merged into a single Frida script (working around frida-java-bridge #218); on a live process each script runs as its own job so toggling one hook-set never blinks the others.
• Session & Script Lifecycle: Automatic detection of session detachment and script destruction — running jobs are marked as errored so callers can react immediately.
• afrim Tool Integration: Utilize the afrim tool to check for existing Frida server installations on your device and ensure you are always running the latest version.

The project was inspired by Frida-Python-Binding.

Install

Just install it via pip:

python3 -m pip install AndroidFridaManager

This will install the afrim-command to your system.

Usage

In order to easily install the latest frida-server version to your Android device just run the following command:

$ afrim 

In order to check only if frida-server is running invoke it with the -r-parameter:

$ afrim -r

API Usage

In order to install and run Frida on your Android device use the FridaManager-API:

from AndroidFridaManager import FridaManager
...
afm_obj = FridaManager(is_remote=False, socket="ip:port", verbose=False, frida_install_dst="/data/local/tmp/")
afm_obj.install_frida_server()
afm_obj.run_frida_server()

For running Frida scripts as jobs use the JobManager-API:

from AndroidFridaManager import JobManager
...
app_package = "net.classwindexampleyear.bookseapiececountry"
frida_script_path = "./frida_script1.js"
job_manager = JobManager()
job_manager.setup_frida_session(app_package, myAwesomeHookingHandler)
job_manager.start_job(frida_script_path, myAwesomeHookingHandler)
print("Running jobs:", job_manager.running_jobs())
job_manager.stop_app_with_closing_frida(app_package)

The setup_frida_session() function accepts a callback function as its second parameter, typically provided by script.on('message', on_message). This initializes the first job and, by default, is the only job where you can activate child_gating and spawn_gating. To execute the job, you must invoke the start_job() function.

For loading multiple hook-sets on a paused spawn (e.g. bypass + tooling scripts), use the Bundle API:

from AndroidFridaManager import JobManager

job_manager = JobManager()
pid = job_manager.spawn_paused("com.example.app")

bundle_id = job_manager.start_bundle(
    scripts=[
        ("root-bypass", root_bypass_source),
        ("ssl-pinning", ssl_pinning_source),
    ],
    custom_hooking_handler_name=my_handler,
)

job_manager.resume_app()

API

FridaManager

FridaManager(is_remote=False, socket="", verbose=False, frida_install_dst="/data/local/tmp/",
             device_serial=None, adb=None)

install_frida_server(dst_dir=None, version=None)
    # version=None auto-matches host frida package version;
    # pass "latest" for newest GitHub release, or a specific version string

run_frida_server(frida_server_path=None)
    # waits until the frida client can actually reach the server (up to 15s)

is_frida_server_running()
stop_frida_server()
remove_frida_server()

get_installed_server_version()       # returns version string from device, or None
list_available_versions(limit=15)    # fetches recent release tags from GitHub

get_connected_devices()              # classmethod — list ADB devices
get_frida_devices()                  # classmethod — list Frida-visible devices

JobManager

JobManager(host="", enable_spawn_gating=False, device_serial=None, adb=None)

# Session setup
setup_frida_session(target_process, frida_callback_function, should_spawn=True, foreground=False)
setup_frida_handler(host="", enable_spawn_gating=False)

# Spawn / attach
spawn(target_process)
spawn_paused(target_process)         # keeps process paused for multi-tool loading
resume_app()                         # resume a paused spawn after hooks are loaded
mark_resumed()                       # sync flag after an external device.resume() call
is_paused()

# Jobs
start_job(frida_script_name, frida_callback_function, job_type="custom",
          display_name=None, hooks_registry=None, priority=50, auto_resume=True)
running_jobs()
stop_jobs()
stop_job_with_id(job_id)

# Bundle API — declarative multi-script coordination
start_bundle(scripts, custom_hooking_handler_name, *, job_type="bundle",
             display_name=None, hooks_registry=None, priority=50, auto_resume=False)
    # scripts: list of (label, source) pairs
    # paused spawn → merged into one script; live → separate jobs
update_bundle(bundle_id, scripts)    # declarative reconcile (not a delta)
stop_bundle(bundle_id)
bundle_clean_labels(bundle_id)       # labels that loaded without error

# Session lifecycle
attach_app(target_process, foreground=False)
reset_session(timeout_per_job=2.0, detach_timeout=2.0)
get_job_by_id(job_id)

# App control
start_android_app(package_name, main_activity=None, extras=None)
stop_app(app_package)
stop_app_with_closing_frida(app_package)
kill_app(pid)
detach_from_app()

# Hook coordination
register_hooks(job_id, hooks)
unregister_hooks(job_id)
check_hook_conflicts(hooks)
get_hook_registry()

# Session info
has_active_session()
get_session_info()
get_running_jobs_info()
get_all_jobs_info()

# setup_frida_handler, setup_frida_session will raise FridaBasedException(Exception). Ensure to handle it.

The JobManager expects a running frida-server on the target device.