v0.19.2

Fixed

• Fixed an issue where an audio source could cause a segmentation fault on application exit if audio was still being written.

v0.19.1

Fixed

• Fixed an audio renderer issue that would cause garbled audio if remote audio was stereo.

• Handle meeting session ID changes which occur once the session has started.

v0.19.0

Added

• Added a new CustomAudioTrack. This new track can be used as an additional custom audio track (i.e. with custom names) or as the main microphone track.

audio_source = CustomAudioSource(...)

audio_track = CustomAudioTrack(audio_source)

client.join("YOUR_DAILY_ROOM", client_settings={
    "inputs": {
        "microphone": {
            "isEnabled": True,
            "settings": {
               "customTrack": {
                   "id": audio_track.id
               }
            }
        }
    }
})

• Added support for CallClient.start_dialout() new fields: displayName, userId, video, codecs, and permissions.

• CallClient.set_audio_renderer() can receive two additional arguments: sample_rate and callback_interval_ms. With sample_rate you can now specify the desired sample rate of the incoming audio data. With callback_interval_ms you can control how often the provided callback is called (with 10ms intervals).

Changed

• CallClient.add_custom_audio_track() and CallClient.update_custom_audio_track() now receive a CustomAudioTrack instead of a CustomAudioSource.

• System certificates are now loaded on macOS, Linux and Windows platforms instead of the embedded Mozilla's root certificates.

Fixed

• Fixed an issue that would cause a blocking virtual microphone to not send any audio in some cases.

v0.18.2

Fixed

• Fixed daily-python install issue on macOS < 15.0.

• Fixed CallClient.remove_custom_track() type hints.

v0.18.1

Fixed

• Fixed an issue that would cause virtual microphones to not call the completion callbacks in some situations.

• Retrying multiple times to open the signalling channel during the join.

• Alternate websocket URI format to avoid connection issues potentially caused by ISPs.

v0.18.0

Added

• It is now possible to create custom audio tracks with CallClient.add_custom_audio_track(). The custom audio tracks need a custom audio source which can be created with CustomAudioSource(). You can also update an existing custom track with a new audio source with CallClient.update_custom_audio_track() or remove and existing custom track with CallClient.remove_custom_audio_track().

• Support the new canReceive permission, which involves:

  • Showing the proper track state (i.e. "off", with "receivePermission" among the "offReasons") when canReceive permission is lacking for a remote track.
  • Resuming receiving remote tracks when previously-revoked canReceive permissions have been restored.

Changed

• Error messages "no subscription for consumer" and "Failed to close consumer" are now logged as debug messages since they can be caused by a remote participant simply disconnecting which is a valid and common case.

• Audio (CallClient.set_audio_renderer()) and video (CallClient.set_video_renderer()) renderers callbacks now receive the audio and video source as the last argument.

v0.17.0

Fixed

• daily-python 0.16.x was compiled with glibc 2.35, which prevented installing it on systems with an older glibc version. This release goes back to glibc 2.28.

v0.16.1

Fixed

• Fixed an issue that was preventing new virtual speakers to get selected.

v0.16.0

Added

• Added support for updating remote participants' canReceive permission via the update_remote_participants() method.

Fixed

• Fixed an issue where the ICE servers configured by the user were not being applied when creating the PeerConnection.

v0.15.0

Added

• Added support to automatically start a recording when joining a room if the start_cloud_recording token property is set to True.

Fixed

• Fixed an issue where reconnection would fail if the meeting token was no longer valid.