[Freeswitch-users] Speech Recognition recording to URL

[mayamatakeshi]

Hi,
out of curiosity, I'm trying to see if I can use application record +
mod_httapi (https://freeswitch.org/confluence/display/FREESWITCH/mod_httapi)
to connect with Google Speech Recognition service (via a proxy app that
will handle HTTP to gRCP convertion).
Just as a test, I'm trying to see if using an inbound event socket I can
inject a DETECTED_SPEECH event in a channel.
I tried something like this and got an +OK from FS:

T 192.168.88.78:60380 -> 192.168.88.74:8021 [AP]
sendevent DETECTED_SPEECH
Unique-ID: a03486b8-781d-430a-ab26-5815485401a8
Call-UUID: a03486b8-781d-430a-ab26-5815485401a8
call_uuid: a03486b8-781d-430a-ab26-5815485401a8
uuid: a03486b8-781d-430a-ab26-5815485401a8
Channel-UUID: a03486b8-781d-430a-ab26-5815485401a8
Speech-Type: begin-speaking

T 192.168.88.74:8021 -> 192.168.88.78:60380 [AP]
Content-Type: command/reply
Reply-Text: +OK 64bdd4aa-9476-4db8-b56c-36a62f6a2238

but nothing got notified to the channel (and no logs about it from FS
running in debug level)

Would this be actually possible?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20200524/f6b1ea2e/attachment.html>