VAPI usage with Trieve

gistfile1.txt

import requests

import os

from dotenv import load_dotenv


#import the VAPI_API_KEY from the .env file
load_dotenv()
VAPI_API_KEY = os.getenv("VAPI_API_KEY")


headers = {
     'Authorization': f'Bearer {VAPI_API_KEY}',
     "Content-Type": "application/json"
 }

# #Creating knowledge base

# data_kb = {
#     "provider" : "trieve",
#     "name": "vapi.ai",
# }

# knowledge_base = requests.post('https://api.vapi.ai/knowledge-base', headers=headers, json=data_kb)

# Create Knowledge Base (POST /knowledge-base)
knowledge_base = requests.post(
  "https://api.vapi.ai/knowledge-base",
  headers={
    "Authorization": "Bearer {VAPI_API_KEY}",
    "Content-Type": "application/json"
  },
  json={
    "provider": "trieve",
    "createPlan": {
      "type": "import",
      "providerId": "{DATASET_ID}"
    },
    "name": "data_kb",
    "searchPlan": {
      "searchType": "bm25"
    }
  },
)









assistant_id = os.getenv("ASSISTANT_ID")
phone_number_id = os.getenv("PHONE_NUMBER_ID")
customer_number = os.getenv("CUSTOMER_NUMBER")

knowledge_base_id = knowledge_base.json()['id']

data_call = {

    "assistant": {

        "transcriber" : {
            "provider": "deepgram",
            "language": "en",
            "model": "nova-2",
        },

        "model": {
            "provider": "openai",
            "model": "gpt-4o",
            "knowledgeBaseId": "{knowledge_base_id}",

            "messages": [
                {
                    "role": "assistant",
                    "content": """You are a friendly voice agent for journaling purposes. 
                    Talk to your user in a friendly way and record everything that the user says. 
                    Keep the conversation hyper personalized by using the previous conversations as a context. 
                    Use the knowledge base for getting context for hyper personalized conversation."""
                }
            ]           
        },

        "voice":{
            "provider": "openai",
            "voiceId": " alloy",
        }
    },    
    
    "phoneNumberId": "{phone_number_id}",
    "customer": {
        "number": "{customer_number}"
    },
}

create_call = requests.post('https://api.vapi.ai/call', headers=headers, json=data_call)

call_id = create_call.json()['id']

# Fetch call details using call_id
call_details_url = f'https://api.vapi.ai/call/{call_id}'


call_details_response = requests.get(call_details_url, headers=headers)

summary = call_details_response.json()['analysis']['summary']






#Adding chunks to the knowledge base
url = "https://api.trieve.ai/api/chunk"


payload = {"chunk_html": "{summary}"}
headers = {
    "Authorization": "{TRIEVE_API_KEY}",
    "TR-Dataset": "{DATASET_ID}",
    "Content-Type": "application/json"
}

requests.request("POST", url, json=payload, headers=headers)