Composer

Overview

Composer is Vapi’s intelligent assistant that helps you build and configure voice AI agents through natural conversation. Instead of manually configuring settings and writing prompts, describe what you want to build and Composer handles the technical setup.

Composer understands voice agent architecture and Vapi’s capabilities. It can create agents, configure phone numbers, set up integrations, troubleshoot issues, and answer questions about Vapi features.

Why use Composer:

• Faster development: Build agents in minutes instead of hours by describing your use case
• Best practices built in: Composer applies Vapi best practices automatically
• Lower barrier to entry: No need to learn every API parameter or configuration option
• Troubleshooting support: Composer can diagnose issues and suggest fixes

Get started

What Composer can do

Capabilities

• Create and configure assistants
• Set up phone numbers
• Configure integrations and webhooks
• Update agent prompts and settings
• Troubleshoot technical issues
• Create and run simulations, and generate a test suite for an assistant
• Answer questions about Vapi features
• Recommend best practices for your use case

Limitations

• Cannot delete resources: Composer cannot delete assistants, tools, phone numbers, or any other resources. This is an intentional safety measure. Use the dashboard sidebar to delete resources manually. Composer can discard your own unpublished assistant or tool draft, which never changes the live resource.
• Cannot access external systems: Composer operates within Vapi and cannot reach your CRM, database, or other third-party systems directly.
• Cannot make business logic decisions: You provide the domain knowledge; Composer handles the technical configuration.
• Cannot test every behavior: Generated suites run in chat, so they can’t check voice-only behavior such as pronunciation. They also skip some handoffs and transfers. Test these with real calls.
• Cannot deploy to production automatically: You control when and how changes go live.

Common use cases

You: "I want to build an agent for my dental office that can schedule
appointments and answer common questions."
Composer: "I'd be happy to help you build a dental office agent! Let me
ask a few questions:
1. What are your office hours?
2. What types of appointments do you offer?
3. Do you want the agent to book directly or collect information
   for your staff to follow up?
4. What are the most common questions patients ask?"
You: "Office hours are Mon-Fri 9am-5pm. We do cleanings, exams, and
consultations. The agent should collect appointment requests and our
staff will confirm. Common questions are about insurance, parking,
and first-visit procedures."
Composer: "I've created your dental office agent. It's configured to:
- Collect appointment requests during business hours
- Answer questions about insurance, parking, and first visits
- Escalate to staff when needed
Would you like me to set up a phone number for testing?"

Generate tests for an assistant

Ask Composer to generate tests for an assistant, for example “Generate a test suite for my dental office agent.” Composer reads the assistant’s prompt and tools and drafts a list of testable requirements. Once you agree on them, it builds a simulation suite that covers them.

After generating a suite, ask Composer to change it in the same conversation:

• Remove simulations: Removed simulations are unlinked from the suite, not deleted.
• Rewrite a simulation: Describe the change, for example “make the caller refuse to give a date of birth.”
• Retry failed simulations: When some simulations couldn’t be generated, Composer offers to try them again.

Composer saves the agreed requirements with the suite, and every edit works from them. If you change the requirements later, Composer builds a new suite. The existing suite doesn’t change. It remains in your account until you delete it from the dashboard.

Safety features

Composer includes safeguards to prevent accidental or irreversible changes to your account. You control how much Composer can do on its own with per-conversation approval modes.

Approval modes

You choose when Composer pauses for your approval, and the setting persists for that conversation:

• Ask before every change: Composer requests approval for all creates, updates, deletes, runs, and publishes.
• Ask before updates, deletes, runs, and publishing: Composer creates new resources on its own, but asks before changing, deleting, running, or publishing anything that already exists.
• Bypass approvals: Composer applies changes without pausing.
• Custom: set permissions for each resource and action individually.

Approval enforcement always happens on the server, and read operations never require approval.

Deletion

Composer does not delete resources. Delete assistants, tools, phone numbers, squads, and files from the dashboard yourself. When Composer removes a simulation from a suite it generated, the simulation is unlinked from the suite, not deleted.

How the approval flow works

You: "Change my agent's voice to sound more energetic"
Composer: [Proposes update]
UI shows: "Update Resource, updating voice settings on assistant xyz"
  [Approve] [Deny]
You: [Clicks Approve]
Composer: "Done! I've updated the voice settings. Test it out and let me
know if you'd like to adjust further."

Key details about approvals:

• Tokens expire after 10 minutes: if you don’t respond in time, Composer re-proposes the change.
• Each approval is specific: approval tokens are cryptographically bound to the exact change being made, so multiple changes each require individual approval.
• Read operations don’t require approval: Composer can freely read and list your resources.

Tips for best results

Be specific about your use case

Provide context about your industry, audience, and requirements upfront.

Less effective: "I need an agent"
More effective: "I need an agent for my e-commerce store that can track
orders, answer product questions, and handle returns"

Iterate in stages

Build a basic working agent first, then layer on advanced features:

1. Phase 1: Basic conversation flow
2. Phase 2: Add integrations (calendar, CRM, knowledge base)
3. Phase 3: Add advanced features (custom voices, complex routing)

Handle one task at a time

Keep conversations focused on a single objective for best results.

Less effective: "Set up my phone number, fix the transfer issue, add
Spanish support, and integrate with my CRM"
More effective: "First, let's set up my phone number"
[Complete that task]
"Great! Now can you help me fix the transfer issue?"

Ask Composer to explain

If you’re unsure about Vapi features or want to learn the API, ask directly:

• “What’s the best way to handle voicemail?”
• “Should I use function calling or server URL for my integration?”
• “Can you show me the API call you’re making to create this agent?”

Next steps

Now that you know how Composer works:

• Assistants quickstart: Learn the fundamentals of building voice agents manually
• Prompting guide: Write effective system prompts for your agents
• Tools overview: Understand the tools and integrations available to your agents
• Phone numbers: Set up phone numbers for your agents