Vapi
Vapi · Voice AI Agent Platform
A developer platform for building, testing, and deploying voice AI agents that handle phone and voice conversations at scale. Best for engineering teams that need to automate customer-facing calls — support, qualification, scheduling — with full control over telephony, voice, and backend logic.
Best for
Developers and enterprises building custom voice agents that need telephony, real-time orchestration, and enterprise security and compliance
Not ideal for
Teams wanting a ready-made no-code support agent, or use cases that are purely text-based
Who it's for
Developers and enterprises building production voice AI agents for support, sales, and scheduling
Capabilities
- Build, test, and deploy voice agents from a single API-first platform
- Real-time call orchestration and monitoring across the voice stack
- Configurable voice characteristics, conversation flows, and backend integrations
- Telephony support for inbound and outbound calls
- Enterprise security: SSO, OAuth2, RBAC, plus SOC 2 and PCI compliance; HIPAA available as a paid add-on
- Squads: orchestrate multiple specialized assistants with context-preserving transfers
- 100+ languages via configurable STT/TTS providers, with automatic language detection
- AI guardrails designed to reduce hallucinations in live calls
- SDKs and documented API for embedding into your own systems
Limitations
- Developer-oriented — production agents require engineering work, not a no-code setup
- Usage-based costs ($0.05/min hosting plus provider costs) scale with call volume and need modeling at scale
- Conversation quality depends on the models and voice providers you connect
- Language quality varies by language and by the STT/TTS provider you choose
Use cases
- Automating inbound customer support phone lines
- Outbound lead qualification and follow-up calls
- Appointment scheduling and reminders by voice
- Routing and triaging calls before handing off to humans
- Building branded voice assistants embedded in products
Our take
Vapi is the build-it-yourself option in voice AI. Where Decagon sells a finished support product and ElevenLabs Agents bundles its own best-in-class voice, Vapi gives you a provider-agnostic orchestration layer and gets out of the way. That is exactly right for engineering teams that want to own the agent logic, telephony, and data flow end to end — and exactly wrong for teams that just want a working support line without writing code. Choose Vapi when control and flexibility matter more than turnkey speed.
Who should use it
Engineering teams and enterprises that want to build and own custom voice agents — choosing their own models and voices, wiring in their own backends, and meeting enterprise security and compliance requirements.
Who should skip it
Teams that want a ready-made, no-code support agent, have no engineering capacity, or only need text-based chat rather than voice.
Strengths
- Full control over the voice stack — voice, conversation flow, telephony, and backend logic
- Real-time call orchestration and monitoring across providers
- Provider-agnostic: bring your own LLM and TTS/STT providers
- Enterprise security and compliance (SSO, RBAC, SOC 2, PCI; HIPAA as an add-on)
- 100+ languages through configurable speech providers
- API-first with SDKs for embedding into your own systems
Weaknesses
- Requires engineering work — not a no-code, prebuilt agent
- Usage-based costs grow with call volume and need modeling at scale
- Conversation quality depends on the models and voice providers you connect
- HIPAA eligibility costs an extra $2,000/month
Vapi pricing
Free
$0 + usage
- $5 free credits
- 4 concurrent calls
- 1 phone number
- 14-day retention
Core
$29/month + usage
Billed monthly
- 10 concurrent calls
- 5 phone numbers
- 30-day retention
Pro
$999/month minimum
Billed monthly
- Minimum equals 10% of hosting fees
- 30 concurrent calls
- 99% uptime SLA, Slack support
- 180-day retention
Premier
Custom
- 99.9% uptime SLA
- 1-hour P0 response
Note: Vapi hosting is $0.05/min (prepaid, usage-based); STT, LLM, TTS, and telephony provider costs are billed on top. Add-ons: extra concurrency at $10/line/month, and HIPAA eligibility with a BAA at $2,000/month.
Technical specs
Modalities
Voice, Text
API pricing
Usage-based: $0.05/min Vapi hosting plus provider costs (STT, LLM, TTS, telephony)
Available models
Where Vapi excels
Automating an inbound support phone line
Telephony plus real-time orchestration lets a Vapi agent answer calls, resolve common requests, and hand off to humans when needed.
Outbound qualification and reminder calls
Programmatic outbound calling with backend integrations lets teams qualify leads or confirm appointments at scale.
Embedding a branded voice assistant in a product
API-first design and SDKs let developers ship a custom voice experience inside their own app rather than a third-party widget.
Vapi vs. competitors
Vapi vs. ElevenLabs Agents
ElevenLabs Agents bundles its own best-in-class voice and is quicker to deploy; Vapi is a provider-agnostic orchestration layer, better when you want to choose your own voice and model providers.
Vapi vs. Decagon
Decagon is a managed, buy-it customer support product; Vapi is build-it infrastructure — pick Vapi to own the agent logic, Decagon for a ready-made multi-channel support solution.
Frequently asked questions
What is Vapi?
Vapi is an API-first platform that provides the infrastructure for voice AI agents — speech, telephony, and real-time orchestration — so developers can build, test, and deploy voice agents for support, sales, and scheduling without assembling the stack themselves.
How does Vapi pricing work?
Vapi charges $0.05 per minute for hosting, with STT, LLM, TTS, and telephony provider costs billed on top. Plans are Free ($5 in credits), Core ($29/month), Pro ($999/month minimum, with a 99% uptime SLA), and custom Premier. HIPAA is a $2,000/month add-on. Model your call volume before committing.
What languages does Vapi support?
Language support depends on the speech providers you configure. Supported providers cover 100+ languages — for example Deepgram (100+), Google STT (125+), and Azure TTS (140+) — with automatic language detection; Vapi Voices covers 40+. Quality varies by language and provider.
Can I use my own models and voices with Vapi?
Yes. Vapi is provider-agnostic — you configure the language model and TTS/STT providers, the voice characteristics, and the conversation flow, while Vapi handles orchestration and telephony.
How does Vapi compare to ElevenLabs Agents?
ElevenLabs Agents bundles ElevenLabs' own high-quality voice and is faster to stand up; Vapi is a provider-agnostic orchestration layer you assemble, trading turnkey voice for flexibility and control over the full stack.
Integrations & fit
Alternatives to consider
ElevenLabs
Teams building voice-first agents where natural speech quality and multilingual coverage matter, especially for phone support and receptionists
FreemiumDecagon
Large consumer brands automating high-volume, multi-channel customer support with on-brand, measurable AI agents
CustomAbout Vapi
Vapi is an API-first platform that provides the infrastructure for voice AI agents so teams don't have to assemble the speech, telephony, and orchestration layers themselves. You configure the voice, conversation flow, and backend integrations, and Vapi handles real-time call orchestration and monitoring across the stack. It is aimed at developers and enterprises building production voice experiences rather than a no-code support product — the platform exposes SDKs, a documented API, and webhook-style hooks into your own systems. Enterprise capabilities include SSO, OAuth2, and role-based access controls, SOC 2 and PCI compliance, HIPAA eligibility as a paid add-on, uptime SLAs on higher plans, and AI guardrails intended to reduce hallucinations. Squads let you orchestrate several specialized assistants with context-preserving transfers. It is built for high call volumes with low-latency responses and supports 100+ languages through configurable speech-to-text and text-to-speech providers (such as Deepgram, Google, and Azure), with automatic language detection. Pricing is usage-based: Vapi charges $0.05 per minute for hosting on top of provider costs (STT, LLM, TTS, telephony), with a Free plan that includes $5 in credits, Core at $29/month, Pro at a $999/month minimum, and custom Premier — cost-effective to start but worth modeling at high call volumes. It fits teams that want to own the voice agent experience end to end, and is overkill for anyone who just needs a prebuilt chatbot.
Are you the founder? Claim this listing →