Add Anannas AI LLM service docs

[Haleshot]

Drafting it for now till pipecat-ai/pipecat#2848 gets merged.

[markbackman]

Your repo looks great! Do you have a quick video you can share of it working. It can just be ~30 seconds using Loom or similar. With that, I can approve. Nice work! 🙌

[Haleshot]

Hi @markbackman, apologies for the delayed response (not sure how I missed the notification for this ping 😓); re: the short demo video, got it "working" but I'm facing a weird error with the messages not showing up on the UI in the conversation bar. On talking to the bot, I get a response back (audio) as well as in the logs on the terminal. The requests also get sent to our anannas dashboard (as seen from the end of the video where we show the log).

Configured the required API keys in the .env file (Anannas, AZURE (Speech and Region API/endpoints) and DEEPGRAM).

Attached demo:

compressed-demo.mp4

Any help in this regard would be greatly appreciated.

PS: Apologies for posting debugging queries on a thread which should otherwise be a simple merge. Hope you don't mind.

[Haleshot]

Hi @markbackman, just thought I'd ping back on this thread again 😅

[markbackman]

@Haleshot, sorry for the delay. Transcripts displaying in the UI requires an RTVIProcessor in your Pipeline and an RTVIObserver. You can see that in use here:
https://github.com/pipecat-ai/pipecat/blob/main/examples/foundational/38b-smart-turn-local.py#L82-L103

[markbackman]

LGTM! Apologies for the delayed review 🙇