janus-client 0.7.3

Janus WebRTC gateway Python async client.

Janus WebRTC gateway Python async client.

---

Install

pip install janus-client

Requires Python >=3.8 <3.14

---

Description

Easily send and share WebRTC media through Janus WebRTC server.

This client is using aiortc for WebRTC communication and subsequently PyAV for media stack.

✅ Features ✅

• Connect to Janus server using:
  • Websocket
  • HTTP
• Authentication with shared static secret (API key) and/or stored token
• Support Admin/Monitor API:
  • Generic requests
  • Configuration related requests
  • Token related requests
• Support Janus plugins:
  • EchoTest plugin
  • VideoCall plugin (Please refer to eg_videocall_in.py and eg_videocall_out.py)
  • VideoRoom plugin
• Simple interface
• Minimum dependency
• Extendable Janus transport

---

Examples

Simple Connect And Disconnect

import asyncio
from janus_client import JanusSession, JanusEchoTestPlugin, JanusVideoRoomPlugin

# Protocol will be derived from base_url
base_url = "wss://janusmy.josephgetmyip.com/janusbasews/janus"
# OR
base_url = "https://janusmy.josephgetmyip.com/janusbase/janus"

session = JanusSession(base_url=base_url)

plugin_handle = JanusEchoTestPlugin()

# Attach to Janus session
await plugin_handle.attach(session=session)

# Destroy plugin handle
await plugin_handle.destroy()

This will create a plugin handle and then destroy it.

Notice that we don't need to call connect or disconnect explicitly. It's managed internally.

Make Video Calls

import asyncio
from janus_client import JanusSession, JanusVideoCallPlugin
from aiortc.contrib.media import MediaPlayer, MediaRecorder

async def main():
    # Create session
    session = JanusSession(
        base_url="wss://janusmy.josephgetmyip.com/janusbasews/janus",
    )

    # Create plugin
    plugin_handle = JanusVideoCallPlugin()

    # Attach to Janus session
    await plugin_handle.attach(session=session)

    # Prepare username and media stream
    username = "testusernamein"
    username_out = "testusernameout"

    player = MediaPlayer(
        "desktop",
        format="gdigrab",
        options={
            "video_size": "640x480",
            "framerate": "30",
            "offset_x": "20",
            "offset_y": "30",
        },
    )
    recorder = MediaRecorder("./videocall_record_out.mp4")

    # Register myself as testusernameout
    result = await plugin_handle.register(username=username_out)

    # Call testusernamein
    result = await plugin_handle.call(
        username=username, player=player, recorder=recorder
    )

    # Wait awhile then hangup
    await asyncio.sleep(30)

    result = await plugin_handle.hangup()

    # Destroy plugin
    await plugin_handle.destroy()

    # Destroy session
    await session.destroy()


if __name__ == "__main__":
    try:
        asyncio.run(main())
    except KeyboardInterrupt:
        pass

This example will register to the VideoCall plugin using username testusernameout. It will then call the user registered using the username testusernamein.

A portion of the screen will be captured and sent in the call media stream. The incoming media stream will be saved into videocall_record_out.mp4 file.

Documentation

The project documentation is built with Material for MkDocs and deployed to GitHub Pages.

Documentation Development

Setup

Install development dependencies:

poetry install --with dev

Local Development

To serve the documentation locally with live reload:

poetry run mkdocs serve

The documentation will be available at http://127.0.0.1:8000/

Building Documentation

To build the documentation for production:

poetry run python -W ignore::DeprecationWarning:mkdocs_autorefs -m mkdocs build --clean --strict

The built documentation will be in the site/ directory.

Important: Always use the --strict flag when building documentation to catch warnings as errors. This ensures documentation quality and prevents deployment of documentation with issues.

For local development without strict mode:

poetry run mkdocs build

Documentation Structure

• docs/index.md - Main documentation page
• docs/session.md - Session API documentation
• docs/plugins.md - Plugin API documentation
• docs/transport.md - Transport API documentation
• mkdocs.yml - MkDocs configuration file

Experiments

FFmpeg support for VideoRoom plugin has now been moved to experiments folder, together with GStreamer support.

TODO 2. Run UT in all supported python versions. 3. Implement text room plugin 4. At documentations, move session, plugin, and transport to be linked pages at the left side. Then move class references to another "References" page. 4. Make websocket connection able to clean up properly when encounter exception. Test using .\tests\test_admin.py::TestTransportWebsocket::test_add_and_remove_token

Use following command to run unit tests and see all logs:

hatch test .\tests\test_plugin.py::TestTransportHttp::test_plugin_echotest_create -- -s --log-cli-level=INFO --full-trace
hatch test -- -s --log-cli-level=INFO --full-trace -- tests