ToolsCompaniesAbout

Vapi

by VapiYC W21No ratings yet

Developer platform for building and deploying voice AI agents

Visit vapi.ai

Linked from vapi.ai · checked 29 Oct

Tool type
Voice AI agent developer platform
Made for
Voice-agent developers and product teams
Runs on
Cloud API, dashboard and CLI
Calls run on Vapi infrastructure; telephony and network transport are provided by Vapi or third-party carriers.
Last checked
29 Oct 2026
3 product sources

From @Vapi_AI

Recent important posts · features

Open on X

What it does

Vapi takes in a voice conversation over phone calls or web sessions and produces a live voice agent that transcribes speech, runs a language model to decide responses, and speaks replies back. Builders configure the agent through a dashboard, a CLI, or an API, choosing the transcriber, model and voice components along with tools that connect to their own systems.

People typically start from a template or a Model Intelligence preset in the dashboard, publish a version, test it with simulations, and then review calls and quality alerts. Assistants handle single-prompt use cases such as support or lead qualification, while Squads coordinate multiple specialized assistants with context-preserving transfers. Configuration can be kept as files in git for review and promotion between environments.

Vapi suits developers and product teams building inbound or outbound phone agents or embedded web voice features, and enterprises that need SSO, RBAC and contractual SLAs. Access is through the web dashboard, API, and CLI, with sign-up at dashboard.vapi.ai. Usage is billed per minute on top of provider costs for transcription, model and voice, and telephony is charged by the provider. Compliance terms such as HIPAA are offered as add-ons or plan features rather than defaults.

What it can do

  • Build voice agents with Assistants using a single system prompt, tools and structured outputs
  • Coordinate multiple specialized assistants with Squads and context-preserving transfers
  • Make and receive phone calls on phone numbers, and embed voice calls in web applications
  • Connect agents to APIs and systems through tool integration
  • Manage configuration as code via CLI and git with simulation testing before deployment
  • Monitor calls and surface quality issues with observability and alerts

Works with

  • Deepgram
  • OpenAI
  • ElevenLabs
  • Anthropic
  • Google
  • Gladia
  • Twilio
  • Vonage
  • Telnyx

Example workflows

Inbound customer support agent
Illustrative example. Input: a support knowledge base and a phone number. Steps: create an Assistant from a template, attach tools for knowledge lookup and escalation, publish, and run simulations. Output: an inbound agent that answers calls and escalates to a human when needed.
Appointment scheduling with Squads
Illustrative example. Input: scheduling rules and an external booking API. Steps: create specialized assistants for triage and booking, link them in a Squad, connect the booking API as a tool, and test transfers. Output: a multi-assistant flow that routes callers and confirms appointments.

Before you choose

  • Per-minute usage excludes provider costs for transcription, model and voice; these are billed at provider rates shown on the pricing page.
  • Telephony transport fees are charged by the carrier (for example Twilio, Vonage, Telnyx) and vary by provider.
  • HIPAA eligibility, data retention, and SSO depend on plan or add-ons, and compliance claims should be confirmed in contract terms.

Listed under

Get started

Create an account and start building; a Docs quickstart is also available.

  1. Create an account at dashboard.vapi.ai and select a template or Model Intelligence preset.
  2. Choose transcriber, model and voice providers, then attach a phone number or embed the web voice widget.
  3. Publish a version, run Simulations to test behavior, and review calls and alerts before going live.
Create an account

Plans and prices

Prices in USD

Vapi hosting is $0.05/minute, plus model, speech and carrier costs. The $5 signup credit is one-time. Optional Core is $29/month; Pro costs 10% of hosting fees with a $999/month minimum. Extra concurrency, organizations and eligible compliance arrangements have separate charges.

Usage-based (No Success Package)$0.05 per minuteVapi hosting only. Model, transcription, voice and carrier costs are additional.
  • $5 free credits to start
  • 2 organizations and 4 concurrent calls
Core$29/moBilled monthly as a Success Package on top of usage.
  • 5 included phone numbers and 10 concurrent calls
  • Email support with 2 business day response SLA
Pro$999/mo10% of the Vapi hosting fee, subject to a $999 monthly minimum; usage and provider costs are additional.
  • 10 organizations and 30 concurrent calls
  • RBAC, ZDR, and 99% uptime SLA

Source: vapi.ai/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

Ratings and reviews

No ratings for Vapi yet.

Be the first to rate it

What’s changed

Free account

See what’s changed at Vapi

Launches, price moves and pivots, each with a date and a source. Free with an account.

Signals

Free account

See Vapi’s signals

Launch traction, hiring and release activity, tracked over time. Free with an account.

Data and security

Free account

See how Vapi handles your data

Training on your files, compliance, hosting and the models behind it. Free with an account.

Compare with similar tools

Free account

Compare Vapi side by side

Prices, platforms and features against similar tools, with sources. Free with an account.

Where this comes from

How we research
Company website
vapi.ai: product, pricing and policy pages
Hacker News
Launch posts and discussion
YC directory mirror
Batch and company name
Social accounts
Accounts the company links, and its posts on X
Human review
Checked by a person
knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.