Vapi

What it's good for

• 1

  Inbound call handling: an agent that answers, qualifies, and routes callers

• 2

  Outbound lead-generation calls that follow a script and book meetings

• 3

  Appointment booking and reminders over the phone

• 4

  After-hours coverage so no call goes unanswered

• 5

  Voice front-ends for support that deflect common questions before a human is needed

Pricing

• Platform fee

  $0.05/min on top of your STT + LLM + TTS + telephony

  Usage-based

• All-in typical

  Roughly $0.15 to $0.33/min once every piece is added

  $0.15/min

• Concurrency

  10 concurrent calls included, extra lines $10 each

  $10/line/mo

• Enterprise

  Volume pricing and support

  Custom

How to use it

Define an assistant with a system prompt, then choose your STT, LLM, and TTS providers (Vapi is model-agnostic, so you pick the voice and brain). Connect a phone number and set up inbound or outbound flows. Give the agent tools (function calls) so it can look up a booking, check availability, or hand off to a human mid-call. Test in the dashboard, watch the latency, and tune the pieces until the conversation feels natural. Budget on the real all-in per-minute rate, not just the platform fee, and add concurrent lines as call volume grows.

Pros & cons

Frequently asked questions

• How much does Vapi cost?

  Vapi charges a $0.05/min platform fee, but that is on top of the speech-to-text, LLM, text-to-speech, and telephony you plug in. The realistic all-in cost is about $0.15 to $0.33 per minute depending on which providers you choose. Ten concurrent calls are included, extra lines are $10 each per month, and Enterprise pricing is custom.

• Vapi vs Retell AI: which is better?

  They are close competitors for building phone voice agents. Vapi is very developer-first and model-agnostic, giving you fine control over the STT, LLM, and TTS stack. Retell AI has a slightly lower headline voice-engine rate and a higher free concurrency allowance. If you want maximum control over every piece, Vapi fits; if you want a marginally cheaper per-minute rate and more free lines, compare Retell AI closely.

• Is Vapi free?

  No, Vapi is a paid, usage-based platform. There is no free tier for production calls. You pay the $0.05/min platform fee plus the underlying STT, LLM, TTS, and telephony costs, which brings the real rate to roughly $0.15 to $0.33 per minute. You can test in the dashboard, but live phone calls are billed.

• What is Vapi best for?

  Developers building voice AI agents that make or take phone calls: inbound handling, outbound lead-gen calls, booking, and after-hours coverage. Its model-agnostic design suits teams that want to pick their own voice and LLM rather than accept a fixed stack.

• Can Vapi handle multiple calls at once?

  Yes. Ten concurrent calls are included by default, and you add capacity by buying more lines at $10 each per month. For high call volume you would scale concurrency accordingly or move to an Enterprise plan for volume pricing.

More

• Vapi docs
• Vapi pricing

Alternatives to Vapi

Other tools we'd consider for the same job.

• Retell AI

  Voice-agent platform to build, test, and deploy phone agents. Per-minute pricing with generous free concurrency

• Lindy

  No-code agent builder with 1,600+ integrations. The pragmatic answer when LangGraph is overkill but Zapier can't express the logic

Other Agents tools

• Dify

  Source-available, self-hostable platform for LLM apps and agents. Visual workflows, RAG, and agents in one package

• Retell AI

  Voice-agent platform to build, test, and deploy phone agents. Per-minute pricing with generous free concurrency

• n8n

  Open-source workflow automation with first-class LLM + vector-DB nodes. Self-host or use their cloud

• LangGraph

  Code-level framework for building multi-step, stateful LLM agents. The serious engineer's choice for agent orchestration