Skip to content

Getting Started

Overview

Choose the path that matches your use case:

All paths start with the same first two steps.

Step 1 - Create an API Key

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/auth/create-key" \
  -H "Content-Type: application/json" \
  -d '{
    "user_name": "Admin User",
    "user_email": "admin@example.com"
  }'

Save the api_key from the response. All subsequent requests require Authorization: Bearer <api_key>.

Step 2 - Create an Assistant

Assistants support two execution modes:

  • pipeline (default): OpenAI realtime handles STT+LLM and a separate TTS provider speaks output.
  • realtime: Gemini realtime handles STT+LLM+TTS in one model.

Both modes support assistant_interaction_config.speaks_first=true. When enabled, the assistant sends an opening response using assistant_start_instruction (or its default greeting if omitted).

Example A: Create Assistant in pipeline mode

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/assistant/create" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_name": "Support Bot",
    "assistant_description": "Customer support agent",
    "assistant_prompt": "You are a helpful support agent.",
    "assistant_start_instruction": "Hello, thanks for calling support. How can I help you today?",
    "assistant_llm_mode": "pipeline",
    "assistant_tts_model": "mistral",
    "assistant_tts_config": {
      "voice_id": "your_mistral_voice_id"
    },
    "assistant_interaction_config": {
      "speaks_first": true,
      "filler_words": false,
      "silence_reprompts": true
    }
  }'

Example B: Create Assistant in realtime mode

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/assistant/create" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_name": "Gemini Voice Bot",
    "assistant_description": "Realtime conversational assistant",
    "assistant_prompt": "You are a helpful voice assistant.",
    "assistant_start_instruction": "Hi, you're connected to Gemini Voice Bot. How can I assist you today?",
    "assistant_llm_mode": "realtime",
    "assistant_llm_config": {
      "provider": "gemini",
      "model": "gemini-3.1-flash-live-preview",
      "voice": "Puck"
    },
    "assistant_interaction_config": {
      "speaks_first": true,
      "silence_reprompts": true
    }
  }'

Save the assistant_id from the response.

Behavior note: - pipeline mode sends the opening response through the pipeline path (LLM + configured TTS). - realtime mode sends the opening response through the realtime conversation path.

Step 3 (Optional) - Give Your Assistant Tools

Tools let the assistant call external APIs or return static data during a conversation. You can skip this step and add tools later.

Create a tool

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/tool/create" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "tool_name": "get_order_status",
    "tool_description": "Look up the current status of a customer order by order ID",
    "tool_parameters": [
      {
        "name": "order_id",
        "type": "string",
        "description": "The order ID to look up",
        "required": true
      }
    ],
    "tool_execution_type": "webhook",
    "tool_execution_config": {
      "url": "https://your-api.com/orders/status",
      "timeout": 5
    }
  }'

Save the tool_id from the response.

Attach the tool to your assistant

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/tool/attach/ASSISTANT_ID" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "tool_ids": ["TOOL_ID"]
  }'

The assistant will now use this tool whenever the conversation requires it. See Tools for full details.

Path A: Outbound SIP Call

A1 - Create a SIP Trunk

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/sip/create-outbound-trunk" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "trunk_name": "My Twilio Trunk",
    "trunk_address": "your-twilio-sip-domain.sip.twilio.com",
    "trunk_numbers": ["+15550100000"],
    "trunk_auth_username": "your_twilio_username",
    "trunk_auth_password": "your_twilio_password",
    "trunk_type": "twilio"
  }'

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/sip/create-outbound-trunk" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "trunk_name": "My Exotel Trunk",
    "trunk_type": "exotel",
    "trunk_config": {
      "exotel_number": "+918044319240"
    }
  }'

Save the trunk_id (format: ST_...) from the response.

A2 - Trigger the Call

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/call/outbound" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_id": "ASSISTANT_ID",
    "trunk_id": "TRUNK_ID",
    "to_number": "+15550100000",
    "call_service": "twilio"
  }'

A 200 OK response means the call was placed successfully.

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/call/outbound" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_id": "ASSISTANT_ID",
    "trunk_id": "TRUNK_ID",
    "to_number": "+918044319240",
    "call_service": "exotel"
  }'

A 202 Accepted response means call setup has started. The final outcome (answered, busy, no_answer, etc.) is delivered asynchronously via the end-call webhook configured on the assistant (assistant_end_call_url).

Save the room_name from the response to correlate with webhook events and call logs.

Path B: Web Call

Web calls use WebRTC directly - no SIP trunk required. The client joins a LiveKit room in the browser or mobile app.

B1 - Generate a Token

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/web_call/get_token" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_id": "ASSISTANT_ID"
  }'

The response contains a token and room_name. Pass the token to the LiveKit client SDK to join the room.

B2 - Connect with LiveKit SDK

import { Room } from "livekit-client";

const room = new Room();
await room.connect("wss://your-livekit-server", token);

The assistant joins automatically once the room is created. Both audio and typed text input (lk.chat topic) are supported.

Path C: Inbound Call

Inbound calls (Exotel only) route incoming calls to an assistant based on the dialed number.

C1 - Assign an Inbound Number

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/inbound/assign" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_id": "ASSISTANT_ID",
    "service": "exotel",
    "inbound_config": {
      "phone_number": "+918044319240"
    }
  }'

When someone dials +918044319240, the assistant is automatically dispatched into a LiveKit room.

See Inbound Calls and Inbound Context Strategies for full details.

End-to-End Example: Outbound Call with a Tool and Webhook

This walkthrough covers every step from zero to receiving a post-call webhook with transcripts.

1. Create your API key

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/auth/create-key" \
  -H "Content-Type: application/json" \
  -d '{"user_name": "Demo User", "user_email": "demo@example.com"}'

Copy the api_key from the response. All commands below use it as YOUR_API_KEY.

2. Create an assistant with an end-call webhook

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/assistant/create" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_name": "Order Support",
    "assistant_description": "Helps customers check order status",
    "assistant_prompt": "You are a support agent for Acme Corp. Help the customer with their order. The customer name is {{name}}.",
    "assistant_llm_mode": "pipeline",
    "assistant_tts_model": "cartesia",
    "assistant_tts_config": {"voice_id": "a167e0f3-df7e-4277-976b-be2f952fa275"},
    "assistant_interaction_config": {"speaks_first": true, "silence_reprompts": true},
    "assistant_end_call_enabled": true,
    "assistant_end_call_trigger_phrase": "end the call",
    "assistant_end_call_agent_message": "Thank you for calling, goodbye!",
    "assistant_end_call_url": "https://your-server.com/webhook/call-ended"
  }'

Note the assistant_id.

3. Create a tool and attach it

# Create the tool
curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/tool/create" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "tool_name": "check_order",
    "tool_description": "Look up order status by order ID. Use when the customer asks about an order.",
    "tool_parameters": [{"name": "order_id", "type": "string", "required": true}],
    "tool_execution_type": "webhook",
    "tool_execution_config": {"url": "https://your-api.com/orders/lookup"}
  }'

# Attach it to the assistant
curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/tool/attach/ASSISTANT_ID" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"tool_ids": ["TOOL_ID"]}'

4. Create a SIP trunk

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/sip/create-outbound-trunk" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "trunk_name": "My Twilio Trunk",
    "trunk_address": "your-sip-domain.sip.twilio.com",
    "trunk_numbers": ["+15550100000"],
    "trunk_auth_username": "your_username",
    "trunk_auth_password": "your_password",
    "trunk_type": "twilio"
  }'

Note the trunk_id.

5. Trigger the call

curl -X POST "https://api-livekit-vyom.indusnettechnologies.com/call/outbound" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "assistant_id": "ASSISTANT_ID",
    "trunk_id": "TRUNK_ID",
    "to_number": "+15550200000",
    "call_service": "twilio",
    "metadata": {"name": "John Doe"}
  }'

The assistant will call the number, greet "John Doe" using the placeholder, and use the check_order tool if the customer asks about an order.

6. Receive the webhook

When the call ends, your assistant_end_call_url receives a POST with the full call record including transcripts, duration, recording URL, and token usage. See End Call Webhook for the complete payload schema.

Next Steps