Ankit Gordhandas

Pipecat has been instrumental for us at Pear - we use AI communication to communicate with patients. We started using pipecat before pipecat cloud was a thing; they haven't stopped showing love to their open source repo in spite of the cloud product (it helps that the cloud product is also powered by the open source code). We have had no issues, even when we had to switch providers – fortunately pipecat provides integrations with all major players in the space. Their docs are great and customer service is the best. I have been on video calls with their CTO at midnight multiple times, and have always left with a satisfactory result.

sometimes docs can get out of date with frequent updates. Par for the course, and cost of moving fast that I am willing to swallow

At the time, pipecat was farther along. Kwin (Founder/CEO) had a bunch of tweets that showed that he was really into this space. @rajiv_ayyangaralso suggested we use pipecat, so here we are.

Yes - that's what it was built for

Most major ones. We have tried elevenlabs, deepgram and gladia

Not very difficult. Pipecat cloud makes it really easy. We self host though, and most of the scaling problems are solved on the infra level (think: AWS) rather than the pipecat level, which is to say that pipecat has made it really easy.