Skip to content

artubss/SKILLS-CLAUDE-CODE / favorites/create-assistant

create-assistant

Unverified

Create and configure Vapi voice AI assistants with models, voices, transcribers, tools, hooks, and advanced settings. Use when building voice agents, phone bots, customer support assistants, or any conversational AI that handles phone or web calls.

Vapi Assistant Creation

Create fully configured voice AI assistants using the Vapi API. Assistants combine a language model, voice, and transcriber to handle real-time phone and web conversations.

Setup: Ensure VAPI_API_KEY is set. See the setup-api-key skill if needed.

Quick Start

cURL

curl -X POST https://api.vapi.ai/assistant \
  -H "Authorization: Bearer $VAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Support Assistant",
    "firstMessage": "Hello! How can I help you today?",
    "model": {
      "provider": "openai",
      "model": "gpt-4.1",
      "messages": [
        {
          "role": "system",
          "content": "You are a friendly phone support assistant. Keep responses concise and under 30 words."
        }
      ]
    },
    "voice": {
      "provider": "vapi",
      "voiceId": "Elliot"
    },
    "transcriber": {
      "provider": "deepgram",
      "model": "nova-3",
      "language": "en"
    }
  }'

TypeScript (Server SDK)

import { VapiClient } from "@vapi-ai/server-sdk";

const vapi = new VapiClient({ token: process.env.VAPI_API_KEY! });

const assistant = await vapi.assistants.create({
  name: "Support Assistant",
  firstMessage: "Hello! How can I help you today?",
  model: {
    provider: "openai",
    model: "gpt-4.1",
    messages: [
      {
        role: "system",
        content: "You are a friendly phone support assistant. Keep responses concise and under 30 words.",
      },
    ],
  },
  voice: {
    provider: "vapi",
    voiceId: "Elliot",
  },
  transcriber: {
    provider: "deepgram",
    model: "nova-3",
    language: "en",
  },
});

console.log("Assistant created:", assistant.id);

Python

import requests
import os

response = requests.post(
    "https://api.vapi.ai/assistant",
    headers={
        "Authorization": f"Bearer {os.environ['VAPI_API_KEY']}",
        "Content-Type": "application/json",
    },
    json={
        "name": "Support Assistant",
        "firstMessage": "Hello! How can I help you today?",
        "model": {
            "provider": "openai",
            "model": "gpt-4.1",
            "messages": [
                {
                    "role": "system",
                    "content": "You are a friendly phone support assistant. Keep responses concise and under 30 words.",
                }
            ],
        },
        "voice": {"provider": "vapi", "voiceId": "Elliot"},
        "transcriber": {"provider": "deepgram", "model": "nova-3", "language": "en"},
    },
)

assistant = response.json()
print(f"Assistant created: {assistant['id']}")

Core Configuration

Model (required)

The language model powering the assistant's intelligence.

ProviderModelsNotes
openaigpt-4o, gpt-4o-mini, gpt-4-turboMost popular, best tool calling
anthropicclaude-3-5-sonnet-20241022, claude-3-5-haiku-20241022Strong reasoning
googlegemini-1.5-pro, gemini-1.5-flashMultimodal capable
groqllama-3.1-70b-versatile, llama-3.1-8b-instantUltra-fast inference
deepinframeta-llama/Meta-Llama-3.1-70B-InstructOpen-source models
openrouterVariousAccess 100+ models
perplexityllama-3.1-sonar-large-128k-onlineWeb-connected
together-aiVarious open-sourceCost-effective
{
  "model": {
    "provider": "openai",
    "model": "gpt-4.1",
    "temperature": 0.7,
    "maxTokens": 1000,
    "messages": [
      {
        "role": "system",
        "content": "Your system prompt here. Define the assistant's personality, rules, and behavior."
      }
    ]
  }
}

Voice

The text-to-speech voice for the assistant.

ProviderPopular VoicesNotes
vapiElliot, Lily, Rohan, Paola, KianVapi's optimized voices, lowest latency
11labsUse voice IDs from ElevenLabsHigh quality, many voices
playhtUse voice IDs from PlayHTExpressive voices
cartesiaUse voice IDs from CartesiaFast, high quality
openaialloy, echo, fable, onyx, nova, shimmerOpenAI TTS voices
azureAzure voice namesEnterprise-grade
deepgramaura-asteria-en, aura-luna-enLow latency
rime-aiUse voice IDs from RimeSpecialized voices
{
  "voice": {
    "provider": "vapi",
    "voiceId": "Elliot"
  }
}

Transcriber

The speech-to-text engine for understanding callers.

ProviderModelsNotes
deepgramnova-3, nova-2Fastest, most accurate
googlelatest_long, latest_shortGoogle Cloud STT
gladiafast, accurateEuropean provider
assembly-aibest, nanoHigh accuracy
speechmaticsVariousEnterprise STT
talkscriberDefaultSpecialized
{
  "transcriber": {
    "provider": "deepgram",
    "model": "nova-3",
    "language": "en",
    "keywords": ["Vapi:2", "AI:1"]
  }
}

The keywords field boosts recognition of specific words (word:boost format, boost 1-10).

Behavior Configuration

First Message

{
  "firstMessage": "Hello! Thanks for calling Acme Corp. How can I help you today?",
  "firstMessageMode": "assistant-speaks-first"
}

firstMessageMode options:

  • "assistant-speaks-first" — Assistant greets immediately (default)
  • "assistant-waits-for-user" — Assistant waits for caller to speak first
  • "assistant-speaks-first-with-model-generated-message" — LLM generates the greeting

Background Sound

{
  "backgroundSound": "office"
}

Options: "off", "office", "static"

Backchanneling

Enable natural conversational acknowledgments ("uh-huh", "I see"):

{
  "backgroundDenoisingEnabled": true,
  "backchannelingEnabled": true
}

HIPAA Compliance

{
  "hipaaEnabled": true
}

When enabled, Vapi won't store call recordings or transcripts.

Adding Tools

Attach tools so the assistant can take actions during calls.

Using saved tool IDs

{
  "model": {
    "provider": "openai",
    "model": "gpt-4.1",
    "toolIds": ["tool-id-1", "tool-id-2"],
    "messages": [{"role": "system", "content": "..."}]
  }
}

Inline tool definition

{
  "model": {
    "provider": "openai",
    "model": "gpt-4.1",
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "check_availability",
          "description": "Check appointment availability for a given date",
          "parameters": {
            "type": "object",
            "properties": {
              "date": {
                "type": "string",
                "description": "Date in YYYY-MM-DD format"
              }
            },
            "required": ["date"]
          }
        },
        "server": {
          "url": "https://your-server.com/api/tools"
        }
      }
    ],
    "messages": [{"role": "system", "content": "..."}]
  }
}

Hooks

Automate actions when specific call events occur. See hooks reference for details.

{
  "hooks": [
    {
      "on": "customer.speech.timeout",
      "options": {
        "timeoutSeconds": 10,
        "triggerMaxCount": 3
      },
      "do": [
        {
          "type": "say",
          "exact": "Are you still there?"
        }
      ]
    },
    {
      "on": "call.ending",
      "filters": [
        {
          "type": "oneOf",
          "key": "call.endedReason",
          "oneOf": ["pipeline-error"]
        }
      ],
      "do": [
        {
          "type": "tool",
          "tool": {
            "type": "transferCall",
            "destinations": [
              {
                "type": "number",
                "number": "+1234567890"
              }
            ]
          }
        }
      ]
    }
  ]
}

Managing Assistants

List

curl https://api.vapi.ai/assistant \
  -H "Authorization: Bearer $VAPI_API_KEY"

Get

curl https://api.vapi.ai/assistant/{id} \
  -H "Authorization: Bearer $VAPI_API_KEY"

Update

curl -X PATCH https://api.vapi.ai/assistant/{id} \
  -H "Authorization: Bearer $VAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "firstMessage": "Updated greeting!"
  }'

Delete

curl -X DELETE https://api.vapi.ai/assistant/{id} \
  -H "Authorization: Bearer $VAPI_API_KEY"

Common Patterns

Customer Support Agent

{
  "name": "Customer Support",
  "firstMessage": "Thank you for calling! How can I assist you today?",
  "model": {
    "provider": "openai",
    "model": "gpt-4.1",
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful customer support agent for Acme Corp. Be empathetic, concise, and solution-oriented. If you cannot resolve an issue, offer to transfer to a human agent. Keep responses under 30 words."
      }
    ]
  },
  "voice": { "provider": "vapi", "voiceId": "Lily" },
  "transcriber": { "provider": "deepgram", "model": "nova-3", "language": "en" },
  "backchannelingEnabled": true
}

Appointment Scheduler

{
  "name": "Appointment Scheduler",
  "firstMessage": "Hi there! I can help you schedule an appointment. What date works best for you?",
  "model": {
    "provider": "openai",
    "model": "gpt-4.1",
    "messages": [
      {
        "role": "system",
        "content": "You are an appointment scheduling assistant. Collect the customer's preferred date, time, and service type. Confirm details before booking. Be friendly and efficient."
      }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "book_appointment",
          "description": "Book an appointment for the given date, time, and service",
          "parameters": {
            "type": "object",
            "properties": {
              "date": { "type": "string", "description": "YYYY-MM-DD" },
              "time": { "type": "string", "description": "HH:MM in 24h format" },
              "service": { "type": "string", "description": "Type of service" },
              "name": { "type": "string", "description": "Customer name" }
            },
            "required": ["date", "time", "service", "name"]
          }
        },
        "server": { "url": "https://your-server.com/api/book" }
      }
    ]
  },
  "voice": { "provider": "vapi", "voiceId": "Paola" },
  "transcriber": { "provider": "deepgram", "model": "nova-3", "language": "en" }
}

Multilingual Agent

{
  "name": "Multilingual Support",
  "firstMessage": "Hello! How can I help you? / Hola! Como puedo ayudarte?",
  "model": {
    "provider": "openai",
    "model": "gpt-4.1",
    "messages": [
      {
        "role": "system",
        "content": "You are a multilingual support assistant. Detect the caller's language and respond in the same language. You support English and Spanish."
      }
    ]
  },
  "voice": { "provider": "vapi", "voiceId": "Paola" },
  "transcriber": { "provider": "deepgram", "model": "nova-3", "language": "multi" }
}

References

Additional Resources

The Vapi documentation MCP server (vapi-docs) provides the full Vapi knowledge base (for example searchDocs when that MCP is connected).

Configure MCP in this repository (for example .mcp.json for Claude Code). If it is not already present, run:

claude mcp add vapi-docs -- npx -y mcp-remote https://docs.vapi.ai/_mcp/server
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 464kUpdated Claude CodeCodexVerifiedInstall
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 464kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 318kUpdated Claude CodeCodexUnverifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 318kUpdated Claude CodeCodexUnverifiedInstall