Voice Feature Guide

Learn how to use and configure voice capabilities for your Agent5ive agents.

Overview

The Realtime Voice feature allows your agents to listen and respond using natural speech, creating a more intuitive and accessible interaction experience. This feature leverages OpenAI's new GPT-realtime model to provide high-quality voice interactions with support for MCP tools and agent-to-agent communication.

Standard Tools Now Supported in Voice

Voice conversations now support standard tools (Web Search, Maps, News, etc.) in addition to MCP tools and agent-to-agent communication. This provides full tool functionality in voice mode.

What is Realtime Voice?

Realtime Voice allows your agent to have spoken conversations with users using OpenAI's advanced GPT-realtime model. Instead of typing messages back and forth, users can:

Speak to the agent - Users can talk naturally using their microphone
Hear the agent's responses - The agent responds with a human-like voice
See text transcripts - All spoken words are also shown as text
Use standard tools in voice - All standard tools (Web Search, Maps, News, etc.) now work in voice conversations
Use MCP tools in voice - Experimental MCP tools can be accessed during voice conversations
Communicate with other agents - Voice agents can call other agents as tools

This creates a more natural and accessible way to interact with your agent, similar to talking with a virtual assistant like Siri or Alexa, but with all the specialized knowledge and capabilities of your custom agent.

Tool Approval in Voice Conversations

Voice conversations support the same Tool Approval Policy System as other deployment types. This means you can control how MCP tools and agent-to-agent communication require user approval during voice interactions.

Configurable Approval Options:

Auto-approve all tools - MCP tools and agent tools execute automatically without user confirmation during voice conversations
Require approval for all tools - Users must verbally or manually confirm each tool usage before execution
Custom rules per tool - Configure specific approval requirements for individual MCP tools or agent tools (e.g., auto-approve read-only tools, require approval for actions that modify data)

Tool approval policies help you balance convenience with control in voice interactions. For example, you might auto-approve informational tools (like searching a knowledge base) while requiring approval for actions that create or modify data (like creating tasks or sending emails).

Learn More About Tool Approval

For detailed information on configuring tool approval policies, including examples and best practices, visit the Tool Approval Policy System section in the Tool Integration guide.

OpenAI API Key Required

This feature requires you to provide your own OpenAI API key, which is like a special password that lets your agent use OpenAI's voice technology. You'll need to create an account with OpenAI and get this key (we'll show you how). Your key is stored securely and is only used for your agent's voice capabilities.

Getting an OpenAI API Key

To use the Realtime Voice feature, you'll need an OpenAI API key. An API key is like a special password that allows your agent to use OpenAI's voice technology. You'll also need to have credits (or a payment method) in your OpenAI account to pay for the voice service. Follow these steps to get your API key:

Create an OpenAI account

If you don't already have one, sign up for an account at https://platform.openai.com/signup

Navigate to API keys

Once logged in, go to the API section and select 'API keys' from the left sidebar.

Create a new API key

Click on the 'Create new secret key' button, provide a name for your key (e.g., 'Agent5ive Voice'), and create the key.

Copy your API key

Copy the generated API key immediately and store it securely. Note that OpenAI will only show this key once.

Important Security Notes

Never share your API key publicly or commit it to version control systems
Agent5ive securely encrypts your API key in our database
Your key is only used for processing voice interactions for your agents
You can revoke access at any time by deleting the key from your OpenAI dashboard

Adding Credits to Your OpenAI Account

To use the voice features, you'll need to ensure your OpenAI account has sufficient credits. New accounts typically receive some free credits, but you'll likely need to add a payment method for continued use.

Access your OpenAI account settings

Add a payment method

Add a valid credit card or other supported payment method to your account.

Set usage limits (recommended)

Configure usage limits to prevent unexpected charges. You can set monthly spending caps in the 'Usage limits' section.

Add credits (prepaid)

If available for your account type, you can prepay for credits to better manage your spending.

Usage Costs

Using voice features will cost money through your OpenAI account. The cost is based on how much time your users spend talking with your agent. New OpenAI accounts usually get some free credits to start with. Check the OpenAI pricing page for current rates. You can set spending limits in your OpenAI account to control costs.

Continue to Agent Creation

Documentation

Guides

Tutorials

Help