I have used this API to try prototype automatic transcription but not with Kamailio however since you are having trouble with the API and not Kamailio I can help you out even though this is unrelated to SR-users mailing list. You need to first get the API working on your own development computer. I suggest installing https://cloud.google.com/sdk/ and then following the sample guide on recognition using the CLI here https://cloud.google.com/speech/docs/getting-started You should also familiarize yourself with ffmpeg or sox to put your audio in the very particular format their API requires. If the audio is sampled at 8khz do not expect great results. I ultimately went from IBM's API because their transcription accuracy is better. Results vary greatly based on clarity of the original audio.