Get started

Vapi for AI agents

Vapi builds and deploys voice AI agents for phone calls, used by teams that want a natural-sounding assistant to handle calls without a live operator. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.

Connect VapiSee the 38 actions ↓
AI and modelsAPI keyMCP + RESTvapi.ai
AUDIT LOG · VAPIPOLICY: acme-support
09:41:07 · claude · u_8f2read
vapi.list_chats
List Chats✓ allowed · 212ms
09:41:08 · claude · u_8f2read
vapi.get_call
Get Call✓ allowed · 164ms
09:41:09 · claude · u_8f2write
vapi.create_eval
Create Eval✓ approved · approved by user
09:41:12 · claude · u_8f2destructive
vapi.delete_call
Delete Call✕ blocked · policy: deny
EVERY VAPI CALL, ON THE RECORD
01 · USE CASES

What agents do in Vapi.

01

List recent calls

Pull the call log with outcomes as a read-only report.

02

Create a new assistant

Set up a voice assistant with its script and configuration.

03

Confirm before deleting a phone number

Check with the user before deleting a phone number, since any assistant using it would stop receiving calls.

02 · ACTIONS

38 Vapi actions, graded by risk.

Every Vapi action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.

read

10

Look things up. Allowed by default.

  • vapi.get_call
    Get Call
  • vapi.get_file
    Get File
  • vapi.list_chats
    List Chats
  • vapi.list_evals
    List Evals
  • vapi.get_assistant
    Get Assistant
  • vapi.get_monitoring
    List Monitoring Policies
  • vapi.list_assistants
    List Assistants
  • vapi.list_phone_numbers
    List Phone Numbers
  • vapi.get_structured_outputs
    List Structured Outputs
  • vapi.list_provider_resources
    List Provider Resources

write

23

Create and change things. Allow, or ask the user first.

  • vapi.create_eval
    Create Eval
  • vapi.create_policy
    Create Monitoring Policy
  • vapi.create_assistant
    Create Assistant
  • vapi.create_scorecard
    Create Scorecard
  • vapi.create_open_ai_chat
    Create OpenAI Chat
  • vapi.update_phone_number
    Update Phone Number
  • vapi.create_analytics_query
    Create Analytics Queries
  • vapi.create_provider_resource
    Create Provider Resource
  • vapi.upload_file
    Upload File
  • vapi.eval_controller_get
    Get Eval
  • vapi.patch_insight_insight
    Update Insight
  • vapi.eval_controller_update
    Update Eval
  • vapi.tool_controller_update
    Update Tool
  • vapi.call_controller_find_all
    List Calls
  • vapi.chat_controller_get_chat
    Get Chat
  • vapi.tool_controller_find_one
    Get Tool
+ 7 MORE

destructive

5

Delete, cancel or archive. Ask first, or deny outright.

  • vapi.delete_call
    Delete Call
  • vapi.delete_eval
    Delete Eval
  • vapi.delete_phone_number
    Delete Phone Number
  • vapi.eval_controller_remove_run
    Delete Eval Run
  • vapi.chat_controller_delete_chat
    Delete Chat
03 · HOW IT WORKS

Vapi in three steps.

  1. 01Your users connect VapiThey add their Vapi api key on Arc0 Connect, under your brand. It goes straight into the vault.
  2. 02You set the rulesReads run, writes like “create Eval” can wait for the user, and “delete Call” can be denied outright.
  3. 03Any agent can actYour agent calls Vapi through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
POLICY.TS
await arc0.policies.set('vapi', {
  read: 'allow',
  write: 'ask',        // create_eval
  destructive: 'deny',  // delete_call
})

# Claude Code: the same connection, one URL
$ claude mcp add --transport http arc0 \
    https://mcp.arc0.ai/u/u_8f2
04 · AUTH AND DATA

How Vapi connects.

Users add their Vapi api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

The same Vapi connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →

AUTH
API key
CREDENTIALS
Per-tenant encrypted vault
MODEL SEES
Results only, never credentials
AUDIT LOG
Every call, on every plan
05 · WORKS WITH

Use Vapi from any agent.

Claude
ChatGPT
Cursor
Codex
VS Code
OpenAI Agents SDK
Claude Agent SDK
Vercel AI SDK
Mastra
LangGraph
06 · RELATED

More ai and models apps.

Perplexity AI9 actionsAffinda119 actionsAgent Mail5 actionsAI/ML API30 actionsAivoov1 actionsAll Images AI7 actionsAltText.ai10 actionsAltTextLab2 actions
ALL AI AND MODELS APPS →
07 · FAQ

Vapi and Arc0, answered.

Q01

Can I use Vapi with Claude, ChatGPT or Cursor?

Yes. Connect Vapi to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Vapi actions you allow.

Q02

How do users connect Vapi?

Users add their Vapi api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

Q03

Which Vapi actions can my agent take?

38 in total: 10 read, 23 write and 5 destructive, such as “create Eval”. Your policies decide which of them each agent may call.

Q04

Can I stop my agent from deleting things in Vapi?

Yes. Actions like “delete Call” are graded destructive. Set destructive actions to deny, or to ask so the user approves each one, and blocked calls still show up on the audit log.

Q05

Can my own backend call Vapi too?

Yes. The same Vapi connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.

Get started

Plug Vapi into your agent.

Your users connect Vapi once, under your brand. Your agent gets 38 actions behind your policies, with every call on the record.

Connect VapiBrowse all apps →

Free to build · MCP + REST · Audit log on every plan