OBS Studio Python Scripting Cheatsheet

Getting started:

• Python
• OBS Studio Scripting

Consider cloning this repo and running examples(they are self contained) in OBS Studio, most of them will operate on existing text soure. Tip: you can create a copy of script, rename it, and add to OBS. So two of identical scripts will be run in parallel with separated namespaces. Also check out issues if you found any error or have a suggestion and discussions for collaboration,ideas and Q&A.

Table of content

• UI
• Property modification
• Additional input
• obs_data
• Print all source settings and filter names
• save settings as json
• Source's and filters with identifier string
• Add source
• Move source
• Duplicate source
• Add filter to source
• Toggle sceneitem visibility
• Set current scene
• Get set order in scene
• Add scene with sources to current scene
• Events
• Program state
• Signals
• Timing (sequential primitives)
• Hotkeys
• Play sound
• Read and write private data from scripts or plugins
• Browser source interaction
• Access source dB volume level
• Get current profile settings via ffi
• Debug
• Security
• Docs and code examples
• Changes between versions
• Links
• Contribute

UI

Preview
S.obs_properties_add_button(props, "button1", "Refresh1:",callback)
S.obs_properties_add_bool(props,"_bool","_bool:")
S.obs_properties_add_int(props,"_int","_int:",1,100,1)
S.obs_properties_add_int_slider(props,"_slider","_slider:",1,100,1)
S.obs_properties_add_text(props, "_text", "_text:", S.OBS_TEXT_DEFAULT)
S.obs_properties_add_color(props,"_color","_color:")
S.obs_properties_add_font(props,"_font","_font:")
S.obs_properties_add_font(props,"_font","_font:")
bool_p = S.obs_properties_add_bool(props, "_obs_bool", "Yes/No"); S.obs_property_set_long_description(bool_p, "Check if yes,else uncheck")

See also :
https://obsproject.com/docs/reference-properties.html#property-object-functions

Property modification

def callback(props, prop, *args, **kwargs):  # pass settings implicitly
    p = S.obs_properties_get(props, "button")
    n = next(counter)
    S.obs_property_set_description(p, f"refresh pressed {n} times")
    return True
...
def script_properties():
    props = S.obs_properties_create()
    b = S.obs_properties_add_button(
        props, "button", "refresh pressed 0 times", refresh_pressed
    )
    S.obs_property_set_modified_callback(b, callback)
    return props

Full source
See also :
https://obsproject.com/docs/reference-properties.html#property-modification-functions

Additional input

def callback(props, prop, settings):
    _number = S.obs_data_get_int(settings, "_int")
    _text_value = S.obs_data_get_string(settings, "_text")
    text_property = S.obs_properties_get(props, "_text")
    if _number > 50:
        eg.data = _text_value + str(_number)
        S.obs_property_set_visible(text_property, True)
        return True
    else:
        eg.data = ""
        S.obs_property_set_visible(text_property, False)
        return True
...

def script_properties():  # ui

    ...
    number = S.obs_properties_add_int(props, "_int", "Number", 1, 100, 1)
    text_value = S.obs_properties_add_text(
        props, "_text", "Additional input:", S.OBS_TEXT_DEFAULT
    )
    S.obs_property_set_visible(text_value, False)
    S.obs_property_set_modified_callback(number, callback)
    ...

Full source

Note: properties share similar structure , in Python, Lua, C. Example C See also :
https://obsproject.com/docs/reference-properties.html#property-modification-functions

obs_data

• obs_data_get_string
• obs_data_get_int
• obs_data_get_double
• obs_data_get_bool
• obs_data_get_obj
• obs_data_get_array

Print all source settings and filter names

source = S.obs_get_source_by_name(self.source_name)
settings = S.obs_source_get_settings(source)
psettings = S.obs_source_get_private_settings(source)
dsettings = S.obs_data_get_defaults(settings)
pdsettings = S.obs_data_get_defaults(psettings)
print("[---------- settings ----------")
print(S.obs_data_get_json(settings))
print("---------- private_settings ----------")
print(S.obs_data_get_json(psettings))
print("---------- default settings for this source type ----------")
print(S.obs_data_get_json(dsettings))
print("---------- default private settings for this source type ----------")
print(S.obs_data_get_json(pdsettings))
... 
print("[--------- filter names --------")
for i in range(filter_count):
    settings = S.obs_data_array_item(filters, i)
    filter_name = S.obs_data_get_string(settings, "name")
    S.obs_data_release(settings)
    print(filter_name)
print(" filter names of %s --------" % self.source_name)

Full source

Save settings as json

p = Path(__file__).absolute()  # current script path
file = p.parent / "saved_settings.json"
try:
    content = S.obs_data_get_json(Data._settings_)
    with open(file, "w") as f:
        f.write(content)
except Exception as e:
    print(e, "cannot write to file")

Full source
See also :
https://obsproject.com/docs/reference-settings.html
https://obsproject.com/docs/scripting.html#getting-the-current-script-s-path

Source's and filters with identifier string

To identify with obs_source_get_unversioned_id , or creating source/filter.

Source's

Name	Source type identifier string
Browser	browser_source
Color Source	color_source
Display Capture	monitor_capture
Game Capture	game_capture
Image	image_source
Image Slide Show	slideshow
Media Source	ffmpeg_source
Text (GDI+)	text_gdiplus
Window Capture	window_capture

Filters

Name	Source type identifier string
Async Delay	async_delay_filter
Chroma Key	chroma_key_filter
Chroma Key V2	chroma_key_filter_v2
Color Correction	color_filter
Color Correction V2	color_filter_v2
Color Key	color_key_filter
Color Key V2	color_key_filter_v2
Color Grade	color_grade_filter
Compressor	compressor_filter
Crop/Pad	crop_filter
Expander	expander_filter
Gain	gain_filter
GPU Delay	gpu_delay_filter
Image Mask/Blend	mask_filter
Invert Polarity	invert_polarity_filter
Limiter	limiter_filter
Luma Key	luma_key_filter
Luma Key V2	luma_key_filter_v2
Mask	mask_filter
Mask V2	mask_filter_v2
Noise Gate	noise_gate_filter
Noise Suppression	noise_suppress_filter
Noise Suppression V2	noise_suppress_filter_v2
Render Delay	gpu_delay
Scaling/Aspect Ratio	scale_filter
Scroll	scroll_filter
Sharpen	sharpness_filter
Sharpen V2	sharpness_filter_v2
Video Delay (Async)	async_delay_filter
VST 2.x Plug-in	vst_filter

Add source

Create source and add it to current scene

S.obs_data_set_string(settings, "text", "The quick brown fox jumps over the lazy dog")
source = S.obs_source_create_private("text_gdiplus", "test_py", settings)
S.obs_scene_add(scene, source)

Full source
See also :
https://obsproject.com/docs/reference-scenes.html

Move source

Get current scene , get source name, move source to location

def __init__(self):
    pos = S.vec2()
    self.location = pos
...
def move_text_source(self):
    current_scene = S.obs_frontend_get_current_scene()
    source = S.obs_get_source_by_name("test_py")
    scene = S.obs_scene_from_source(current_scene)
    scene_item = S.obs_scene_find_source(scene, "test_py")
    if scene_item:
        dx, dy = 10, 10
        print("old values", self.location.x)
        S.obs_sceneitem_get_pos(
            scene_item, self.location
        )  # update to last position if its changed from OBS
        self.location.x += dx
        self.location.y += dy
        print("new values", self.location.x)
        S.obs_sceneitem_set_pos(scene_item, self.location)

Full source

Duplicate source

 def dup(self):
     current_scene = S.obs_scene_from_source(S.obs_frontend_get_current_scene())
     scene_item = S.obs_scene_find_source(current_scene, self.source_name)
     info = S.obs_transform_info()
     crop = S.obs_sceneitem_crop()
     S.obs_sceneitem_get_info(scene_item,info)
     S.obs_sceneitem_get_crop(scene_item,crop)
     duplicate = S.obs_sceneitem_get_source(scene_item)
     duplicated = S.obs_source_duplicate(
         duplicate, "duplicate" + self.source_name, False
     )

     scenes = S.obs_frontend_get_scenes()
     for scene in scenes:
         name = S.obs_source_get_name(scene)
         if name == self.scene_name:
             scene = S.obs_scene_from_source(scene)
             scene_item2 = S.obs_scene_add(scene, duplicated)
             S.obs_sceneitem_set_info(scene_item2,info)
             S.obs_sceneitem_set_crop(scene_item2,crop)

Full source

Add filter to source

Filters are sources,they are not listed in obspython module, you need to know its id from obs_source_info

S.obs_data_set_int(settings, "opacity", 50)
source_color = S.obs_source_create_private(
    "color_filter", "opacity to 50", settings
)
S.obs_source_filter_add(source, source_color)

Full source
See also :
Color correction source
https://obsproject.com/docs/reference-sources.html

Toggle sceneitem visibility

def toggle(self):
    current_scene = S.obs_scene_from_source(S.obs_frontend_get_current_scene())
    scene_item = S.obs_scene_find_source(current_scene, self.source_name)
    boolean = not S.obs_sceneitem_visible(scene_item)
    S.obs_sceneitem_set_visible(scene_item, boolean)

Full source

Set current scene

def set_current_scene(self):
    scenes = S.obs_frontend_get_scenes()
    for scene in scenes:
        name = S.obs_source_get_name(scene)
        if name == self.scene_name:
            S.obs_frontend_set_current_scene(scene)
...
scenes = S.obs_frontend_get_scenes() # Dropdown menu UI
for scene in scenes:
    name = S.obs_source_get_name(scene)
    S.obs_property_list_add_string(p, name, name) 

Full source

Get set order in scene

def get_order(scene_items=None):
    order = list()
    for i, s in enumerate(scene_items):
        source = S.obs_sceneitem_get_source(s)
        name = S.obs_source_get_name(source)
        order.append({"index": i, "name": name, "scene_item": s})
    return order


def reorder():
    current_scene = S.obs_frontend_get_current_scene()
    with scene_ar(current_scene) as scene:
        with scene_enum(scene) as scene_items:
            order = get_order(scene_items)
            # change second index with pre last
            order[1]["index"], order[-2]["index"] = (
                order[-2]["index"],
                order[1]["index"],
            )
            for s in sorted(order, key=lambda i: i["index"]):
                S.obs_sceneitem_set_order_position(s["scene_item"], s["index"])

Full source

Add scene with sources to current scene

def add_random_text_source(scene):
    r = " random text # " + str(randint(0, 10))
    with data_ar() as settings:
        S.obs_data_set_string(settings, "text", f"random text value {r}")
        with source_create_ar("text_ft2_source", f"random text{r}", settings) as source:
            pos = S.vec2()
            pos.x = randint(0, 1920)
            pos.y = randint(0, 1080)
            scene_item = S.obs_scene_add(scene, source)
            S.obs_sceneitem_set_pos(scene_item, pos)

def add_scene_with_sources():
    current_scene_source = S.obs_frontend_get_current_scene()
    with scene_from_source_ar(current_scene_source) as scene_source:
        with scene_create_ar("_nested_scene") as _scene:
            py_scene_source = S.obs_scene_get_source(_scene)

            with scene_from_source_ar(py_scene_source) as scene:
                add_random_text_source(scene)
                add_random_text_source(scene)
                add_random_text_source(scene)

            # add created scene to current scene ( nested scene)
            _scene_source = S.obs_scene_get_source(scene)
            S.obs_scene_add(scene_source, _scene_source)

Note: sometimes OBS crashes if one of such scenes has been deleted.

• Full source

Events

def on_event(event):
    if event == S.OBS_FRONTEND_EVENT_SCENE_CHANGED:
        raise Exception("Triggered when the current scene has changed.")


def script_load(settings):
    S.obs_frontend_add_event_callback(on_event)

Full source
See also:
https://obsproject.com/docs/reference-frontend-api.html#structures-enumerations

Program state

Those functions return true or false :

• S.obs_frontend_preview_program_mode_active()
• S.obs_frontend_replay_buffer_active()
• S.obs_frontend_recording_active()
• S.obs_frontend_recording_paused()
• S.obs_frontend_streaming_active()

Signals

Signals , callbacks , differences from C

Core signals

sh = S.obs_get_signal_handler()
S.signal_handler_connect(sh,"source_create",callback)
def callback(calldata):
    source = S.calldata_source(cd,"source")
    print(S.obs_source_get_name(source))

source_create, source_destroy, source_remove, source_save, source_load, source_activate, source_deactivate, source_show, source_hide, source_rename, source_volume, source_transition_start, source_transition_video_stop, source_transition_stop, channel_change, master_volume, hotkey_layout_change, hotkey_register, hotkey_unregister, hotkey_bindings_changed

https://obsproject.com/docs/reference-core.html#core-obs-signals

Scene signals

def connect_cur_scene():
    source = S.obs_frontend_get_current_scene()
    sh = S.obs_source_get_signal_handler(source)
    S.signal_handler_connect(sh, "item_add", callback)
    S.obs_source_release(source)


def callback(calldata):
    scene_item = S.calldata_sceneitem(calldata, "item")
    #scene = S.calldata_source(cd,"scene") # bad utf symbols 
    scene =  S.obs_sceneitem_get_scene(scene_item)
    name = S.obs_source_get_name
    source = S.obs_sceneitem_get_source
    scene_source = S.obs_scene_get_source
    scene_name = name(scene_source(scene))
    scene_item_name = name(source(scene_item))
    print(f"item {scene_item_name} has been added to scene {scene_name}")

• Full source

item_add, item_remove, reorder, refresh, item_visible, item_locked, item_select, item_deselect, item_transform

https://obsproject.com/docs/reference-scenes.html#scene-signals

Source signals

sh = S.obs_source_get_signal_handler(some_source)
S.signal_handler_connect(sh,"show",callback)
def callback(calldata):
    source = S.calldata_source(cd,"source")
    print("on source show",S.obs_source_get_name(source))

destroy, remove, save, load, activate, deactivate, show, hide, mute, push_to_mute_changed, push_to_mute_delay, push_to_talk_changed, push_to_talk_delay, enable, rename, volume, update_properties, update_flags, audio_sync, audio_mixers, filter_add, filter_remove, reorder_filters, transition_start, transition_video_stop, transition_stop, media_started, media_ended, media_pause, media_play, media_restart, media_stopped, media_next, media_previous

https://obsproject.com/docs/reference-sources.html#source-signals

Output signals

def connect_to_rec():
    sh = S.obs_output_get_signal_handler(S.obs_frontend_get_recording_output())
    S.signal_handler_connect(sh, "pause", callback)

def callback(calldata):
    #out = S.calldata_ptr(calldata, "output") # bad type
    print('output paused')

start, stop, pause, unpause, starting, stopping, activate, deactivate, reconnect, reconnect_success

https://obsproject.com/docs/reference-outputs.html#output-signals

Timing (sequential primitives)

def script_update(settings):
    eg.source_name = S.obs_data_get_string(settings, "source")
    S.timer_remove(eg.update_text)
    if eg.source_name != "":
        S.timer_add(eg.update_text, 1 * 1000)

Full source
Note: each time script updated it's removed first
See also :
Version with globals and only one timer allowed.
https://obsproject.com/docs/scripting.html#script-timers

Thread

def callback(pressed):
    if pressed:
        toggle_thread()


def busy_thread():
    while True:
        if not data.thread_paused:
            sleep(0.02)
            data.status = "active"
            # print to stdoud crashes OBS on exit
        else:
            sleep(0.5)
            data.status = "inactive"


print('Press the "~" to toggle on/off')
hook("OBS_KEY_ASCIITILDE", "id_", callback)
S.timer_add(lambda: print(data.status), 500)
t = threading.Thread(target=busy_thread)
t.start()

• Full source

Hotkeys

This hotkey example will create hotkeys in settings , but you need to bind it manually.

class Hotkey:
    def __init__(self, callback, obs_settings, _id):
        self.obs_data = obs_settings
        self.hotkey_id = S.OBS_INVALID_HOTKEY_ID
        self.hotkey_saved_key = None
        self.callback = callback
        self._id = _id

        self.load_hotkey()
        self.register_hotkey()
        self.save_hotkey()

...

class h:
    htk_copy = None  # this attribute will hold instance of Hotkey

...
h1 = h()
h2 = h()
...
def script_load(settings):
    h1.htk_copy = Hotkey(cb1, settings, "h1_id")
    h2.htk_copy = Hotkey(cb2, settings, "h2_id")


def script_save(settings):
    h1.htk_copy.save_hotkey()
    h2.htk_copy.save_hotkey()

This hotkey example will create hotkeys on fly from json settings , but you need to know internal id.

ID = "htk_id"
JSON_DATA = '{"%s":[{"key":"OBS_KEY_1"}]}' % ID

def on_obs_key_1(pressed):
    if pressed:
        raise Exception("hotkey 1 pressed")

def script_load(settings):
    s = S.obs_data_create_from_json(JSON_DATA)
    a = S.obs_data_get_array(s, ID)
    h = S.obs_hotkey_register_frontend(ID, ID, on_obs_key_1)
    S.obs_hotkey_load(h, a)

Here is how send hotkey to OBS

def send_hotkey(obs_htk_id, key_modifiers=None):
    if key_modifiers:
        shift = key_modifiers.get("shift")
        control = key_modifiers.get("control")
        alt = key_modifiers.get("alt")
        command = key_modifiers.get("command")
    ...
    combo = S.obs_key_combination()
    combo.modifiers = modifiers
    combo.key = S.obs_key_from_name(obs_htk_id)
    ...
    S.obs_hotkey_inject_event(combo, False)
    S.obs_hotkey_inject_event(combo, True)
    S.obs_hotkey_inject_event(combo, False)

• Full source
• Example with global
• Full example with json
• Full example with send hotkey

See also:
https://github.com/obsproject/obs-studio/blob/master/libobs/obs-hotkeys.h obsproject/obs-websocket#595

Play sound

def play_sound():
    ...
    mediaSource = S.obs_source_create_private(
        "ffmpeg_source", "Global Media Source", None
    )
    s = S.obs_data_create()
    S.obs_data_set_string(s, "local_file", script_path() + "alert.mp3")
    S.obs_source_update(mediaSource, s)
    S.obs_source_set_monitoring_type(
        mediaSource, S.OBS_MONITORING_TYPE_MONITOR_AND_OUTPUT
    )
    ...

• Full source

Read and write private data from scripts or plugins

Write in one script

def send_to_private_data(data_type, field, result):
    settings = S.obs_data_create()
    set = getattr(obs, f"obs_data_set_{data_type}")
    set(settings, field, result)
    S.obs_apply_private_data(settings)
    S.obs_data_release(settings)

def write_private_data():
    result = "private value from " + str(__file__) + " " + str(randint(1, 10))
    send_to_private_data("string", "__private__", result)

Read from another

@contextmanager
def p_data_ar(data_type, field):
    settings = S.obs_get_private_data()
    get = getattr(obs, f"obs_data_get_{data_type}")
    try:
        yield get(settings, field)
    finally:
        S.obs_data_release(settings)

def print_private_data():
    with p_data_ar("string", "__private__") as value:
        print(value)

Lua is also supported

local obs = obslua
local settings = S.obs_data_create()
S.obs_data_set_int(settings,"__private__", 7)
S.obs_apply_private_data(settings)
S.obs_data_release(settings)

• Full example read
• Full example write

Browser source interaction

def send_hotkey_to_browser(source, obs_htk_id, key_modifiers=None, key_up=False):

    key = S.obs_key_from_name(obs_htk_id)
    vk = S.obs_key_to_virtual_key(key)
    event = S.obs_key_event()
    event.native_vkey = vk
    event.modifiers = get_modifiers(key_modifiers)
    event.native_modifiers = event.modifiers  # https://doc.qt.io/qt-5/qkeyevent.html
    event.native_scancode = vk
    event.text = ""
    S.obs_source_send_key_click(source, event, key_up)


def press_tab(*p):
    with source_auto_release(G.source_name) as source:
        send_hotkey_to_browser(source, "OBS_KEY_TAB")
        send_hotkey_to_browser(source, "OBS_KEY_TAB", key_up=True)


def press_shift_tab(*p):
    with source_auto_release(G.source_name) as source:
        send_hotkey_to_browser(source, "OBS_KEY_TAB", {"shift": True})
        send_hotkey_to_browser(source, "OBS_KEY_TAB", {"shift": True}, key_up=True)

• Full source

Access source dB volume level

There is FFI ctypes module in Python to wrap native obs lib. However,to run it on GNU/Linux you must start obs with LD_PRELOAD.

ubsdrive3@usbdrive3:~$ LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libpython3.9.so obs

It might be in different directory, use find /usr -name 'libpython*' to find exact location.

volmeter_callback_t = CFUNCTYPE(
    None, c_void_p, POINTER(c_float), POINTER(c_float), POINTER(c_float)
)
...
wrap(
    "obs_volmeter_add_callback",
    None,
    argtypes=[POINTER(Volmeter), volmeter_callback_t, c_void_p],
)
...
@volmeter_callback_t
def volmeter_callback(data, mag, peak, input):
    G.noise = float(peak[0])

• Full source

Get current profile settings via ffi

wrap("obs_frontend_get_profile_config", POINTER(Config), use_lib=G.obsffi_front)
# const char *config_get_string(config_t *config, const char *section,
#                  const char *name)
wrap("config_get_string", c_char_p, argtypes=[POINTER(Config), c_char_p, c_char_p])
wrap("config_num_sections", c_size_t, argtypes=[POINTER(Config)])
wrap("config_get_section", c_char_p, argtypes=[POINTER(Config), c_size_t])


def output_to_stdout():
    cfg = G.obs_frontend_get_profile_config()
    e = lambda x: x.encode("utf-8")
    s = G.config_get_string(cfg, e("SimpleOutput"), e("FilePath"))
    l = G.config_num_sections(cfg)
    for i in range(l):
        tag = G.config_get_section(cfg, c_size_t(i))
    print(s, l)

• Full source

Debug

There is no stdin therefore you can't use pdb , options are:

• using print
• using generated log file (contains information about memory leaks)
• using pycharm remote debugging (localhost)
• using threading
• using vscode attach to the process:
  • Load python extension
  • open script file , pip install debugpy , place debugpy.breakpoint() somewhere
    • Run (F5) select configuration ( Attach using Process ID)
    • select obs (on windows obs64.exe)
    • View select Debug Console (ctrl+shift+y)
• Example debugpy obs

Security

• "Calling home" - see #4044
• On GNU/Linux there is handy program called tcpdump, to run a check against OBS Studio - use this command (it is active on 443 port on start and on end)

# tcpdump -i <your_interface_e_g_wifi_or_wire> 'port 443'

• Avoid using sudo or admin, check hashsums, do backups, do updates, setup a firewall, do hardening, etc...
• There is no confirmation for loading Lua or Python scripts - they can be added/overwritten via .json key
• Also solutions for Python source code obfuscation & loading from shared (compiled) library do exist. Applying that will make it a bit harder to reverse-engineer your private code.
• Legal info link
• Privacy policy - https://obsproject.com/privacy-policy - keyword downlodable software

Docs and code examples

• Generated export.md

contains all variables and functions available in obspython formatted with markdown. Table consist of links to appropriate search terms in OBS Studio repository, links to scripts in obspython and obslua with each script within github code search.gs_* and matrix_* functions exluded from that table. Export names

Changes between versions

• Since OBS Studio 28.0 Beta 1 it is possible to use any python3 version on hardware which supports Qt 6.
• Qt 6 has dropped support for Windows 7 & 8, macOS 10.13 & 10.14, Ubuntu 18.04 and all 32-bit operating systems. As such, OBS will no longer be supported on these platforms.
• Added native support for websocket

Links

• Scripts forum , Github topic obs-scripts , Github topic obs-script
• OBS Studio Repo , obs-scripting-python.c
• Docs , Docs/scripting , Docs/plugins , Docs index
• obspython Gist , Github , grep.app
• obslua Gist , Github , grep.app
• A Python bundle for integration with OBS scripting
• Lua tips and tricks
• Python 3.6.8 , 64 bit installer, for Microsoft Windows

Contribute

Contributions are welcome!