Overview

VoiceAgents is a multi-tenant voice agents platform that lets you deploy AI-powered phone agents in minutes. Configure your agent once, assign a phone number, and start handling inbound and outbound calls — no telephony expertise required.

Every agent runs a real-time STT → LLM → TTS pipeline: speech from the caller is transcribed, sent to your chosen language model, and the response is synthesized back as natural voice — all within milliseconds.

Key Features

Architecture

Each call runs through three stages in real-time:

Create an Account

1. 1Go to the sign-up page and enter your name, email, and a strong password (min 6 chars, one uppercase, one number, one special character).
2. 2After sign-up, you'll be prompted to create your organisation. Enter your company name — this sets up your tenant.
3. 3You'll land on the Dashboard. A free trial credit is applied automatically so you can start making calls right away.

Create Your First Agent

1. 1Navigate to Agents in the sidebar and click New Agent.
2. 2Enter an agent name and write a system prompt describing the agent's role and behaviour.
3. 3Select your STT, LLM, and TTS providers (see Agent Setup for details).
4. 4Click Save. Your agent is ready to handle calls.

Get an API Key

API keys let your applications interact with VoiceAgents programmatically.

1. 1Go to Developer → API Keys in the sidebar.
2. 2Click Create New Key, give it a descriptive name (e.g. "Production App").
3. 3Copy the key immediately — it starts with eq_live_ and is shown only once.

Make Your First API Call

Initiate an outbound call using your API key:

curl -X POST https://your-api-url.com/v1/calls/outbound \
  -H "x-api-key: eq_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "agentId": "your-agent-id",
    "toNumber": "+15551234567",
    "fromNumber": "+15559876543"
  }'

Response:

{
  "callId": "550e8400-e29b-41d4-a716-446655440000",
  "status": "initiated"
}

Use the callId to poll for status, retrieve the transcript, or fetch the recording.

Speech-to-Text (STT)

The STT provider transcribes the caller's audio in real time. Choose the provider and model that best fits your language, accuracy, and latency requirements.

Deepgram Options

Language Model (LLM)

The LLM generates the agent's responses based on the system prompt, conversation history, and any retrieved knowledge base context.

Parameters

Text-to-Speech (TTS)

The TTS provider converts the LLM's text response into speech that the caller hears.

Common Settings

Call Settings

These settings control how the agent behaves during a call.

Response Rate Presets

Authentication

All /v1/* endpoints require an API key passed in the x-api-key header.

curl https://your-api-url.com/v1/agents \
  -H "x-api-key: eq_live_your_key_here"

API keys follow the format:

Keys are exactly 72 characters long. Any key that does not match this exact format is rejected before any database lookup.

Rate Limits

Rate limits are applied per tenant (all API keys for an account share the same limit).

When rate limited, the response includes a Retry-After header and a retryAfter field (seconds) in the body.

Agents

Create Agent — Request Body

{
  "name": "Support Agent",
  "systemPrompt": "You are a helpful support assistant for Acme Corp...",
  "greetingMessage": "Hello! How can I help you today?",
  "sttConfig": { "provider": "deepgram", "model": "nova-3", "language": "en" },
  "llmConfig": { "provider": "openai", "model": "gpt-4.1-mini", "temperature": 0.5 },
  "ttsConfig": { "provider": "openai-tts", "voice": "nova" },
  "maxCallDurationS": 3600,
  "recordCalls": true
}

Calls

Outbound Call — Request Body

{
  "agentId": "550e8400-e29b-41d4-a716-446655440000",
  "toNumber": "+15551234567",
  "fromNumber": "+15559876543",
  "overrides": {
    "systemPrompt": "Optional per-call system prompt override"
  }
}

Call Status Values

Error Codes

All errors return a consistent JSON body:

{
  "error": "not_found",
  "code": "AGENT_NOT_FOUND",
  "message": "Agent xyz not found",
  "statusCode": 404
}