Hello, I am trying to implement an asterisk system - ChatGPT, using TTS and STT so that the AI calls or answers and interacts with a client. I saw Mr. Joshua Colp’s presentation last March 12 2024, where he discussed the topic of transcription and AI with Asterisk and recommended using ARI, but I did not see that he talked about AudioSocket like an option. My question would be: What would be the best way to develop a project like this with Asterisk: ARI or AudioSocket?
In my opinion, start by considering that all documentation and sample code to stream media to an external service are written based on ARI and external media. In the end, if you are not quite sure how to implement real-time transcription, the best decision is to use ARI External media and Google Speech Recognition, as there is already sample code and good documentation available for the process. Regarding AudioSocket, I would say it is one of many methods, similar to ChanSpy, but these methods have the disadvantage of a lack of documentation and sample codes. Unless you know how to do it, go with ARI
Thank you, for your advise. I will use ARI external media
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.