create-assistant

What this skill does

# 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

```bash
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)

```typescript
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

```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.

| Provider | Models | Notes |
|----------|--------|-------|
| `openai` | `gpt-4o`, `gpt-4o-mini`, `gpt-4-turbo` | Most popular, best tool calling |
| `anthropic` | `claude-3-5-sonnet-20241022`, `claude-3-5-haiku-20241022` | Strong reasoning |
| `google` | `gemini-1.5-pro`, `gemini-1.5-flash` | Multimodal capable |
| `groq` | `llama-3.1-70b-versatile`, `llama-3.1-8b-instant` | Ultra-fast inference |
| `deepinfra` | `meta-llama/Meta-Llama-3.1-70B-Instruct` | Open-source models |
| `openrouter` | Various | Access 100+ models |
| `perplexity` | `llama-3.1-sonar-large-128k-online` | Web-connected |
| `together-ai` | Various open-source | Cost-effective |

```json
{
  "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.

| Provider | Popular Voices | Notes |
|----------|---------------|-------|
| `vapi` | `Elliot`, `Lily`, `Rohan`, `Paola`, `Kian` | Vapi's optimized voices, lowest latency |
| `11labs` | Use voice IDs from ElevenLabs | High quality, many voices |
| `playht` | Use voice IDs from PlayHT | Expressive voices |
| `cartesia` | Use voice IDs from Cartesia | Fast, high quality |
| `openai` | `alloy`, `echo`, `fable`, `onyx`, `nova`, `shimmer` | OpenAI TTS voices |
| `azure` | Azure voice names | Enterprise-grade |
| `deepgram` | `aura-asteria-en`, `aura-luna-en` | Low latency |
| `rime-ai` | Use voice IDs from Rime | Specialized voices |

```json
{
  "voice": {
    "provider": "vapi",
    "voiceId": "Elliot"
  }
}
```

### Transcriber

The speech-to-text engine for understanding callers.

| Provider | Models | Notes |
|----------|--------|-------|
| `deepgram` | `nova-3`, `nova-2` | Fastest, most accurate |
| `google` | `latest_long`, `latest_short` | Google Cloud STT |
| `gladia` | `fast`, `accurate` | European provider |
| `assembly-ai` | `best`, `nano` | High accuracy |
| `speechmatics` | Various | Enterprise STT |
| `talkscriber` | Default | Specialized |

```json
{
  "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

```json
{
  "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

```json
{
  "backgroundSound": "office"
}
```

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

### Backchanneling

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

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

### HIPAA Compliance

```json
{
  "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

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

### Inline tool definition

```json
{
  "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](references/hooks.md) for details.

```json
{
  "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

```bash
curl htt