Provides the ability to use OpenAI Speech-To-Text service in the Freeswitch.
Also capable to work with whisperd
UPD 30-12-2024:
seems the issue with noises after playback done (play_and_detect_speech) has been fixed in Freeswitch 1.10.10+ (at least in my case)
<extension name="openai-asr">
<condition field="destination_number" expression="^(3222)$">
<action application="answer"/>
<action application="play_and_detect_speech" data="/tmp/test2.wav detect:openai"/>
<action application="sleep" data="1000"/>
<action application="log" data="CRIT SPEECH_RESULT=${detect_speech_result}"/>
<action application="hangup"/>
</condition>
</extension>