How much does Vapi cost to start?
Vapi's entry price is $0 to start, billed on a per-minute usage basis (Vapi pricing). New accounts get $5 in free credits, so you can build and place test calls before paying anything.
The headline rate is $0.05/min for Vapi hosting. That number is a floor, not a total. Vapi passes model provider costs through at cost with no markup, per its pricing page, so your real per-minute rate stacks several components:
- Hosting: $0.05/min (Vapi)
- Transcription (Deepgram): $0.0095–$0.0099/min
- Intelligence (OpenAI): $0.0077–$0.0452/min
- Voice (ElevenLabs): $0.0146–$0.0238/min
- Transport: Vapi Telephony/SIP, WebSockets and Daily WebRTC are free; Twilio inbound is $0.008/min and outbound $0.014/min
Add the low ends together and a typical call minute lands around $0.08–$0.09 before transport. Use premium models and voices and it can exceed $0.12/min. Vapi is a developer platform — your team builds and maintains the agent — so factor engineering time into total cost as well.
Checked 28 October 2026.
Vapi pricing plans compared
Beyond usage, Vapi sells optional Success Packages that raise limits and support levels. Usage charges apply on top of every tier.
| Plan | Price | What's included |
|---|---|---|
| No Success Package | $0 to start | $5 free credits, 2 organizations, 1 included Vapi phone number, 4 concurrent calls, 14 days raw data retention, full model catalog, peer-to-peer Discord support |
| Core | $29/month | 5 included Vapi phone numbers, 10 concurrent calls, 30 days raw data retention, ZDR included, email support, 2 business day response SLA |
| Pro | $999/mo minimum (10% of Vapi hosting fee) | 10 organizations, 10 included Vapi phone numbers, 30 concurrent calls, 180 days raw data retention, RBAC + ZDR, email + dedicated Slack support, 1 business day response SLA, 99% uptime SLA |
| Premier | Contact Sales | Elevated orgs, concurrency and retention, custom phone numbers, priority support, named support team, 1 hour P0 response SLA, 99.9% uptime SLA, custom RBAC + ZDR + SSO, dedicated SIP available |
Two things to note on the Pro tier. First, it costs 10% of your Vapi hosting fee with a $999/month minimum — so you pay $999 until your hosting spend alone passes $9,990/month. That is steep for smaller teams. Second, the free tier's 4 concurrent calls cap is easy to hit if you run inbound and outbound at the same time.
For most teams testing the waters, the free tier plus usage is the real starting point. Core at $29/month is cheap insurance once you need more than 4 simultaneous calls or email support.
What does one month of Vapi cost at three volumes?
We modelled three monthly volumes using the same assumptions throughout: 3.0 minutes average call length, Vapi Telephony (free transport), and no Success Package. Because Vapi's total rate depends on which providers you pick, we show the hosting-only floor and a realistic configured range using the published provider ranges (Deepgram $0.0095–$0.0099, OpenAI $0.0077–$0.0452, ElevenLabs $0.0146–$0.0238 per minute).
That gives a low end of $0.05 + $0.0095 + $0.0077 + $0.0146 = $0.0818/min, and a high end of $0.05 + $0.0099 + $0.0452 + $0.0238 = $0.1289/min.
| Monthly volume | Minutes | Hosting-only floor ($0.05/min) | Configured range ($0.0818–$0.1289/min) |
|---|---|---|---|
| 100 calls | 300 | 300 × $0.05 = $15 | 300 × $0.0818–$0.1289 = $24.54–$38.67 |
| 300 calls | 900 | 900 × $0.05 = $45 | 900 × $0.0818–$0.1289 = $73.62–$116.01 |
| 1,000 calls | 3,000 | 3,000 × $0.05 = $150 | 3,000 × $0.0818–$0.1289 = $245.40–$386.70 |
Assumptions and caveats:
- These use Vapi's own published provider ranges; your actual model and voice choices set where you land in the range.
- Twilio transport would add $0.008/min inbound or $0.014/min outbound (at 1,000 inbound calls: 3,000 × $0.008 = $24 more).
- Add Core ($29/month) if you need 10 concurrent calls or email support; add Pro's $999/month minimum only at real scale.
- We have not made test calls yet, so treat these as arithmetic from published rates, not measured bills.
For context, a human receptionist costs about $3,167/month at the BLS median annual wage of $38,010 — but a receptionist is a different product from a build-it-yourself voice platform.
Charges outside the plan price
Vapi's pricing has several line items beyond the hosting fee and Success Packages:
- Model provider costs: transcription, intelligence and voice are billed at cost on top of the $0.05/min hosting fee (ranges in the table above).
- Telephony transport: Twilio inbound $0.008/min, outbound $0.014/min. Vapi Telephony/SIP, WebSockets and Daily WebRTC are free.
- HIPAA compliance: a paid add-on at $2,000 — not included by default, even though the site lists HIPAA among its compliance options.
- Extra concurrency: $10/line/month beyond your plan's included concurrent calls.
- Extra organizations: $20/organization/month beyond the included count.
- Pro package scaling: 10% of your Vapi hosting fee once that exceeds the $999/month minimum.
The practical takeaway: quote your all-in per-minute rate before committing, not the $0.05 headline. A team on Twilio with premium voices could pay more than double the hosting floor.
Is there a free way to try Vapi?
Yes. Vapi gives new accounts $5 in free credits with no paid plan required. At the $0.05/min hosting floor, that covers up to 100 hosting minutes — less once model, voice and transport costs stack on, but enough to build an agent and place real test calls.
The free tier is genuinely usable for evaluation: full model catalog, 1 included Vapi phone number, 4 concurrent calls and 14 days of raw data retention. Support is peer-to-peer via Discord, so expect community answers rather than a helpdesk.
Our suggestion: use the credits to run a structured test — book an appointment, handle an angry caller, attempt a transfer — before spending on Core or Pro. Our guide on how to test an AI receptionist has a checklist you can adapt for a voice platform like this.
Cheaper or simpler options than Vapi
Vapi suits developer teams that want usage-based pricing and full control. If that is not you, three alternatives from our directory:
- Retell AI — also per-minute, but publishes an all-in range of $0.07–$0.31/minute depending on voice, model and telephony. You get $10 in free credits and 20 free concurrent calls at signup, with no contract. Watch for add-ons like branded call ID at $0.10 per outbound call and billing that continues during silence and hold time.
- Voiceflow — a collaborative build platform spanning chat and voice. Pricing is quote-only for businesses, so you must talk to sales; agencies get a free trial with no credit card. Better fit if CX and product teams want to design together rather than hand everything to engineers.
- Synthflow — enterprise-focused, with contracts starting at $30,000 annually and no published per-minute rate (Synthflow pricing). Only worth a look for high-volume operations needing native telephony and contact-center integrations.
If you want a ready-made receptionist with a flat monthly price and no engineering work, none of these platforms fit — browse our AI receptionist category instead. For a like-for-like platform comparison, see AI voice agent platforms or run the numbers in our AI call center savings calculator.