Vapi

What is Vapi?

Vapi is a developer-first platform for building, testing and deploying voice AI agents that make and answer phone calls. It is API-first and model-agnostic: you assemble an agent from your choice of speech-to-text, language model and text-to-speech provider, and Vapi orchestrates the real-time pipeline — turn-taking, interruptions, telephony and function calls — at sub-500ms latency. It is one of the most widely adopted voice-agent stacks among technical teams building custom phone automation. Founded in 2023 by University of Waterloo classmates Jordan Dearsley (CEO) and Nikhil Gupta (CTO), who had previously built Superpowered, a Y Combinator-backed calendar app, Vapi launched publicly in 2024 and has raised $72M total — a $20M Series A led by Bessemer Venture Partners (December 2024) and a $50M Series B led by Peak XV Partners (May 2026, a reported ~$500M valuation per TechCrunch; the round’s own announcement did not disclose a figure).

What does Vapi do?

You define an agent — its system prompt, the voice and model it uses, and the tools or webhooks it can call — then Vapi runs the live conversation over the phone or the web. During a call the platform streams audio to your chosen transcriber, sends the text to the model, speaks the reply through your chosen TTS voice, and manages the hard real-time problems: detecting when the caller has finished speaking, handling interruptions, and keeping latency low. Agents can call external functions mid-conversation to look up data, book appointments or trigger workflows, and can transfer to a human when needed. Vapi supports inbound and outbound calls, imports your own telephony numbers, and can coordinate multiple specialised agents (‘squads’) that hand off to one another. Everything is driven through the API and SDKs, with a dashboard for building, monitoring and reviewing call transcripts and analytics. Because it is bring-your-own-provider, the same agent can be re-pointed at a different model or voice without being rebuilt.

Key features

• Bring-your-own pipeline: Pick your own STT, LLM and TTS providers; Vapi is model-agnostic rather than locking you to one vendor.
• Real-time orchestration: Handles turn-taking, interruptions and telephony at sub-500ms average latency so calls feel natural.
• Tool / function calling: Agents call external functions and webhooks mid-call to fetch data, book appointments or trigger workflows.
• Squads (multi-agent): Coordinate multiple specialised agents that hand off to one another within a single call.

What are Vapi's use cases?

• Inbound call deflection: A support line answers common questions and completes routine tasks autonomously, escalating to a human only when needed.
• Outbound automation: Appointment reminders and lead-qualification calls run at volume, with the agent updating your systems via function calls.

What does Vapi integrate with?

• Model providers — OpenAI, Anthropic, Google
• Speech providers — bring your own STT & TTS
• Telephony — Twilio, Vonage or imported SIP numbers
• Custom tools via function calling & MCP

Why use Vapi?

• Fine-grained control over every layer of the voice stack — transcription, model, voice and telephony.
• Model- and voice-agnostic, so you avoid single-vendor lock-in and can swap providers per agent.
• API-first with enterprise scaling, compliance add-ons (HIPAA, zero-data-retention) and 99.9% uptime for large deployments.

Pros & cons

Vapi pricing

• Build$0.05 / /min
• ScaleCustom

See current pricing on vapi.ai ↗Compare Vapi alternatives →

Vapi review

Vapi is the builder’s choice for voice AI: an API-first, model-agnostic platform that hands developers control over every layer — transcription, model, voice and telephony — with real-time orchestration and sub-500ms latency. Its bring-your-own-provider design means you can chase the best model or cheapest voice without rebuilding the agent, and enterprise features (HIPAA, zero-data-retention, SLAs) let it scale to millions of calls. The honest catch is cost transparency: the $0.05/min headline is only the orchestration fee, and the true all-in rate is usually several times that once you add a model, a good voice and telephony. Pick Vapi if you have engineers who want to build a bespoke voice agent and own the stack; skip it if you want a no-code tool that hides the plumbing.

The developer’s default for custom voice agents — total control at the price of doing the plumbing yourself and watching the stacked per-minute costs.

Frequently asked questions