I am new, but quick moving, want to use VAPI as a first prototype for my AI voice companion. Need him to perform actions consequent to the call based on the transcription. Thought about using the workflow/infrastructure as suggested below. Also I want to have the transcripts stored in my database (not yet build) in best case matched to a unique ID of person the interaction took place with so the call agent can use the information of the last call in the future ones. I am moving quick, want to be done with the prototype for testing by tomorrow. Any help is therefore much appreciated! All in all I want 6 actions to happen - If i need more for my intent let me know - I think VAPI is involved in step 1, 2 & 5. 1. Voice Companion -3 different uses: Provides instructions, transcribes user input, and motivates user to continue 2. Transcript to Database: Sends transcribed content to the database- Via Webhook (?) 3. Webhook to DB - What DB most appropriate? Match unique ID to user - E.g., for V1 with IP - Any other ideas - IN the end product it must be a Login or something I suppose 4. Database Automated Content Analysis: Analyzes stored content depending on use for themes, goals, and actionable items. - Creates summary 5. Action to User: Sents actionable summaries to user (e.g., Mail) adopts goal setting, or saves content for future search. 6. Database to Voice Companion: Feeds insights back to the Voice Companion from VAPI for iterative goal setting and achievement tracking. Open for suggestions for tools helping me to ease the process. Any best practises?

1. use vapi for: prompting (instructions), transcription(deepgram), and motiavte user to continue (prompting/logic) 2. use assistant server url and end of call report server event more: https://docs.vapi.ai/server-url/events#end-of-call-report and for db poc pov use nosql db such as mongodb 3. please give more clarity 4. upto you how you analyse after call checkout openai 5. same create summaries and action items based on analysis and store in db. implement a simple email service to send actionable summaries to users. for the prototype, you can use a service like SendGrid. 6. use Vapi's api to feed past interaction data into new conversations

Thanks made!

@Johannes resolved?

Not yet fully.

any blocker from vapi?