Zoom AI Services: Voice Agent API

Build intelligent voice
agents that handle real conversations

Give users a voice experience that listens, understands, and acts. Voice Agent API combines real-time speech recognition, LLM based natural-language understanding and tool execution into one managed platform so that every interaction reaches the right outcome, whether that is an instant answer or a completed workflow.

Try demo See pricing

Capabilities

One API to get the conversation going

Voice agent

Create a conversational AI agent that understands user intent, responds in natural language, and resolves issues. Integrate the Voice Agent into your application to provide your users a low latency, fluent and natural way to get answers and take actions.

Function calling

Provide the Voice Agent with access your application and business specific information or other information sources through function calls you define. Or use function calls to take actions on behalf of the user.

Performance

The Voice Agent API with industry-leading accuracy at an affordable price

Highest accuracy

The Voice Agent API's speech recognition is powered by Zoom Scribe which is ranked as a top performing model on the Hugging Face Open ASR Leaderboard.

Affordable and predictable per minute cost

Compare Voice Agent's per minute cost against other Voice Agent APIs.

Based on publicly available pricing data as of the date of publication; actual costs may vary.

One foundation

Make your application easy to use with natural conversational interactions

Start with a voice agent and connect it to your knowledge and backend systems to empower your users to use natural conversational interactions to get the information they need and to take action.

Customer support

Handle order-status checks, returns, billing inquiries, and password resets without a queue. The AI agent can help you confirm details, call your CRM, and read back a resolution — or transfer to a live agent with full context.

Appointment scheduling

Let users book, reschedule, or cancel appointments by voice. Connect a scheduling tool and the AI agent can help you check availability, confirm the slot, and send a confirmation — all in one call.

IT help desk & internal support

Triage IT tickets, answer HR policy questions, process facilities requests, and route complex issues — reducing wait times for employees while capturing structured data for every interaction.

Try it

Call the agent. See it think.

Build and coordinate conversational AI voice agents that can listen, reason, and respond across your workflows.

Sign-up or login is required to access the portal.

Try the playground

Developer first

Start conversing in minutes

Simple, well-documented APIs
REST API and OpenAPI spec
Comprehensive error handling
Flexible summary length controls

See API docs

# Start a voice agent session

npx wscat \
  -c wss://api.zoom.us/v2/aiservices/voice-agent/live \
  -s voice-agent.zoom.v2 \
  -H "Authorization: Bearer $YOUR_JWT_TOKEN"

# After "Connected", send:
{ "type": "session.update", "session": { "listen": {}, "think": {}, "speak": {} } }

# Start a voice agent session
import asyncio, json, os, websockets

async def main():
    async with websockets.connect(
        "wss://api.zoom.us/v2/aiservices/voice-agent/live",
        subprotocols=["voice-agent.zoom.v2"],
        additional_headers={"Authorization": f"Bearer {os.environ['YOUR_JWT_TOKEN']}"},
    ) as ws:
        await ws.send(json.dumps({"type": "session.update", "session": {"listen": {}, "think": {}, "speak": {}}}))
        async for message in ws:
            print(json.loads(message))

asyncio.run(main())

// Start a voice agent session
import WebSocket from 'ws';

const ws = new WebSocket('wss://api.zoom.us/v2/aiservices/voice-agent/live', ['voice-agent.zoom.v2'], {
  headers: { Authorization: `Bearer ${process.env.YOUR_JWT_TOKEN}` },
});

ws.on('open', () => ws.send(JSON.stringify({ type: 'session.update', session: { listen: {}, think: {}, speak: {} } })));
ws.on('message', (data) => console.log(JSON.parse(data)));

Pricing

Simple, transparent pricing

Apply prepaid credits to your use of Voice Agent API with transparent rates.

Voice Agent - Standard API

$1.50/hour

Voice Agent API enables developers to build and deploy conversational voice agents that understand natural speech, engage in dynamic conversations, and execute tasks through seamless voice interactions.

Usage is measured based on the total duration of active voice agent sessions in minutes during a given billing period.Get started

Voice Agent - Bring-Your-Own LLM API

$1.08/hour

Voice Agent - Bring Your Own LLM API enables developers to build and deploy conversational voice agents powered by their own large language model, combining Zoom's voice infrastructure with a custom-integrated LLM to understand natural speech, engage in dynamic conversations, and execute tasks through seamless voice interactions.

Usage is measured based on the total duration of active voice agent sessions in minutes during a given billing period.Get started