Vapi logo

Vapi

Developer platform for building phone and web voice agents where you get to choose the speech-to-text, LLM and voice provider for each agent.

Last verified: October 24, 2026
Alternative to: ElevenLabs
Vapi app screenshot

Short description

Vapi is a platform for building AI phone and web voice agents. Instead of making its own voices, it connects the pieces of an AI phone agent: the phone line, speech recognition, the AI model and the voice, so a business can launch an AI receptionist or support line. It is the main alternative to ElevenAgents, and ElevenLabs is one of its voice providers, so you can keep ElevenLabs voices inside it. What changes is the platform running the calls: you pick each provider and pay their cost directly instead of ElevenLabs' bundled agent package.

Features

  • Choose your own providers: Pick separately among LLMs (OpenAI, Anthropic, Gemini, Groq, Perplexity), voices (ElevenLabs, Cartesia, Azure and others) and transcribers (Deepgram, AssemblyAI, Gladia)

  • Sub-500ms latency claim: The vendor's stated average time between a caller finishing and the agent replying, also claimed at enterprise scale

  • Build, test and deploy: Design call flows, test them and ship them, with live call monitoring and call tracking

  • Free built-in phone and web calling: Vapi Telephony/SIP, WebSockets and Daily WebRTC carry no transport charge, or bring your own carrier

  • Pay per call line: Buy extra simultaneous call lines for $10 each instead of moving to a higher plan

  • Enterprise controls: SSO, OAuth2, role-based access, AI guardrails and zero data retention on paid packages

  • Scale claims: 1M+ developers, 2.5M+ agents and 1B calls

Pricing

  • How you pay: A per-minute platform fee, with speech-to-text, LLM, voice and phone charges billed at each provider's cost

  • Free to start: $0, then pay-as-you-go from prepaid credit

  • No package: $0

  • Core: $29/mo

  • Pro: From $999/mo

  • Premier: Custom

  • Per minute: $0.05 platform fee vs ElevenAgents' $0.08, about $0.08-0.13 all-in

  • Extra call lines: $10 per extra simultaneous call line

  • ElevenLabs voices inside Vapi: $0.0146-0.0238/min at cost

  • HIPAA: $2,000/mo add-on

Integrations

  • Platforms: Web dashboard, a REST API for creating agents in code, phone calls over regular phone lines and web calls in the browser. Cloud-hosted only, with no self-hosting listed.

  • Phone carriers and contact centres: Twilio, Telnyx, Vonage, Five9, Genesys and Avaya connect agents to phone numbers and existing call centres

  • CRM and helpdesk: Salesforce, HubSpot, Zendesk, GoHighLevel, Clay and Apollo.io let agents read and update customer and sales records

  • Productivity and automation: Google Calendar, Google Sheets, Notion, Slack, Zapier and Make let agents book meetings, log data and trigger workflows

  • Storage and monitoring: AWS S3, GCP and Langfuse store call data and track agent behaviour

Platforms

  • Web dashboard
  • REST API
  • Phone calls
  • Web calls
❤️

What users love about Vapi

  • A callable phone agent fast: Call state, transcripts and provider connections are already handled, so teams building receptionist, support and sales agents skip the speech and phone plumbing and get a prototype up quickly

  • Swap models per agent: The LLM, transcriber and voice are picked separately, for example keeping an ElevenLabs voice while trying a different LLM

  • Strong API, docs and Web SDK: Embedding voice into other products and connecting other tools is straightforward

  • Fast hands-on support for active teams: Some production teams get quick, detailed fixes from engineers that keep outages short, and find the platform gets faster and cheaper month to month

😕

What users complain about Vapi

  • Real bill far above $0.05/min: LLM, voice and Twilio charges stack on top of the advertised fee, which customers call hidden pricing, and UK users who must import numbers effectively pay twice per call

  • 3-5 second pauses mid-call: Some callers hit long gaps after each turn, and latency grows in complex multi-step flows, well short of the sub-500ms headline

  • Hard to tell what broke: When a call fails, it's unclear whether the transcriber, LLM, voice or carrier caused it, and developers want latency breakdowns, provider fallback and decision traces

  • Dashboard edits fail to save: Assistant edits fail about half the time, settings reset when other fields change, and documented features like DTMF keypad tones didn't work in testing

  • Slow, automated support for small accounts: Week-long waits, mostly bot replies and refunds still pending weeks after being promised. Interruption handling and memory on long calls are also weak.