Vapi offers two pricing options in 2026: A self-serve, pay-as-you-go option and a custom Enterprise option. I reviewed Vapi's current pricing surfaces, provider billing model, and concurrency docs to show what the call rate covers, where extra costs show up, and which setup fits each team.
Vapi AI Pricing Plans: At a Glance
| Plan | Price | Best For | What You Get |
|---|---|---|---|
| Self-Serve | Calls start at $0.05/minute | Developers, pilots, and low-volume production | Usage-based access, 10 default concurrent call slots, bring-your-own provider keys, and add-on concurrency as needed |
| Enterprise | Custom pricing | Production apps, regulated teams, and higher-volume operations | Higher scale, custom concurrency, enterprise reliability, compliance options, and dedicated support |
Vapi also runs a startup grant for eligible voice AI startups. The program lists up to 90,000 free minutes, 7,500 minutes per month for 12 months, dedicated Slack support, and priority feature access. Applications are on hold.
Most teams evaluating Vapi should budget for a modular stack rather than a flat SaaS fee. The platform fee is only one line item.
What to watch: Many third-party writeups still describe older bundles. Vapi's pricing centers on self-serve usage, a custom Enterprise plan, and a separate plan that's currently on hold.
How Vapi AI Pricing Actually Works
Vapi's $0.05/minute headline is the platform fee for calls, not the full cost of a production voice stack. Your real spend also depends on the provider mix, telephony, concurrency, and any enterprise requirements you add.
| Cost Layer | What It Covers | Why It Changes Spend |
|---|---|---|
| Vapi platform fee | Vapi's orchestration/runtime for calls | This is the public headline rate, but not the whole bill |
| Model and speech providers | STT, LLM, and TTS usage | Provider choice can change costs fast |
| Telephony | Phone numbers and carrier/PSTN costs | Customer-facing call traffic adds a separate transport-cost layer |
| Concurrency | 10 default call slots plus reserved capacity | Burst traffic can push you into add-on spend faster than average-minute models suggest |
| Enterprise controls | Volume pricing, support, procurement, and compliance requirements | Costs shift once you need enterprise coverage rather than pure self-serve |
Vapi's Platform Fee Is Only the Start
Public call rate: Vapi's self-serve call pricing starts at $0.05 per minute. That price covers Vapi's orchestration layer for calls, not the full underlying voice stack.
What still sits on top: Most production deployments still need STT, LLM, TTS, telephony, and any added concurrency or enterprise support.
A layer most Vapi pricing estimates miss entirely: The QA and testing budget for keeping voice agents reliable after every model or prompt change.
Provider Costs Still Sit on Top
Bring your own keys: If you validate your own provider keys, Vapi stops billing you for that provider. The provider bills you directly instead.
Why this matters: The headline call rate can look cheap until you map the full provider mix behind each conversation.
Concurrency Changes the Math Fast
Default limit: Every Vapi account includes 10 concurrent call slots by default.
Upgrade trigger: If launches or outbound bursts push past that baseline, you need to reserve more capacity.
Enterprise Changes What You're Buying
The Enterprise plan follows a different motion: Once you move past self-serve, you're buying scale, reliability, support, and compliance coverage, not just a public per-minute rate.
Vapi Pricing Plans Breakdown
Self-Serve: Calls Start at $0.05/Minute
What's included: Usage-based access, 10 default concurrent call slots, and the ability to bring your own provider keys.
Best for: Developers, pilots, and early production teams that want control over the stack.
Pros:
- Low entry point for call volume
- Flexible provider setup
- Simple way to start testing production traffic
Cons:
- Provider costs still sit outside the headline rate
- Telephony and concurrency can raise the real bill
- You still need to model the full voice stack before launch
Enterprise: Custom Pricing
What's included: Custom scale, higher reliability, custom concurrency, volume pricing, enterprise support, and compliance options.
Best for: Teams running voice agents as a real production channel or buying for regulated environments.
Pros:
- Better fit for scale and procurement
- Stronger support and reliability posture
- More room to negotiate around volume and controls
Cons:
- No public list price
- Heavier sales process
- Harder to benchmark from the website alone
Startup Grant: A Separate Program Rather Than a Standard Plan
What it offers: Vapi's startup grant lists up to 90,000 free minutes, 7,500 minutes per month for 12 months, dedicated Slack support, and priority feature access.
Who qualifies: Vapi lists qualifying startups as teams building voice AI products or features, pre-seed to Series A companies with $250k+ in funding, projects focused on voice applications, and teams launching to customers within six months. Meeting those criteria does not guarantee acceptance.
How to apply: Vapi's startup program is a simple application process, with invite-based onboarding and ongoing support. But applications are currently on hold.
Pros:
- Meaningful free usage for prototyping, testing, and early customer traffic
- Slack support and priority feature access can help technical teams move faster
Cons:
- Eligibility is limited by startup stage, funding, use case, and launch timing
- Applications are on hold, so teams can't treat the grant like a guaranteed discount
Where Most Teams Underbudget With Vapi
Most teams underbudget for Vapi when they consider only the platform fee.
Here's what else to take into account:
- Provider mix: The $0.05/minute call rate doesn't include every upstream provider cost.
- Traffic shape: A modest monthly average can still produce expensive peaks if concurrency spikes.
- Telephony: Phone numbers and transport costs add a separate billing layer.
- Enterprise needs: Support, procurement, and compliance can change the economics once the product is live.
- Internal ops: Someone still has to tune prompts, manage providers, and investigate production failures.
- Volume discounts: If your usage is predictable, custom pricing can matter more than the headline self-serve rate.
Which Vapi Pricing Plan Should You Choose?
The right choice depends less on average minutes and more on how much operational work your team wants to carry.
Choose self-serve if: You want to launch fast, control your provider mix, and model the full stack yourself while traffic is still manageable.
Choose Enterprise if: Voice is already a production channel and you need custom concurrency, stronger support, and a formal compliance or procurement motion.
Consider the Startup Grant Program if: You're an eligible voice AI startup, you match Vapi's qualification criteria, and you can wait. Applications are on hold.
Is Vapi Worth the Cost?
Vapi is worth it for technical teams that want runtime control and are comfortable modeling layered costs. It's a weaker fit for buyers who want one flat bill or very low operational overhead.
Vapi is worth it if you: Want control over providers, can manage the full voice stack, and expect engineering to own production quality and cost modeling.
Skip Vapi if you: Want bundled pricing, minimal stack management, or white-glove enterprise coverage from day one without moving into a custom sales process.
Vapi Alternatives & Pricing Comparison
If you're comparing Vapi with other voice agent platforms, the closest publicly available pricing references today come from Retell, Bland, and Synthflow.
| Tool | Starting Price | Best For | Key Pricing Difference |
|---|---|---|---|
| Retell | Pay-as-you-go. AI voice agents start at $0.07-$0.31/min | Teams that want a usage-based developer platform with voice and chat support | Public range bundles more of the agent stack into one usage figure |
| Bland | Start plan is $0.14/min, Build is $0.12/min, Scale is $0.11/min, and Enterprise plans are custom | Teams that want flat per-minute pricing with LLM, STT, TTS, and telephony included | More all-in-one pricing than Vapi's layered model |
| Synthflow | $0 to start on the pay-as-you-go plan. Enterprise pricing is custom. | Teams that want a builder-oriented PAYG model with calculator-based cost planning | Voice engine, LLM, telephony, and concurrency are priced as separate levers |
How Vapi and Cekura Work Together
Vapi and Cekura don't solve the same problem, so this isn't a winner-loser matchup. Vapi is the orchestration layer. It helps you build, deploy, and run voice agents.
Cekura is the QA layer. It adds pre-production simulations, infrastructure testing, production call QA, and security testing on top of a live voice stack.
Cekura offers native integrations that work out of the box for Retell, VAPI, ElevenLabs, LiveKit, Pipecat, Bland, and more. You don't rebuild anything. You add a testing and voice observability layer on top of what you already have.
Plus, it's SOC 2-, HIPAA-, and GDPR-compliant for transcript redaction, role-based access, and audit trails.
That pairing matters when you need to catch workflow regressions, interruption failures, latency issues, or adversarial behavior before they hit customers.
Lindy is a good example of this pattern: Cekura helped the team stress-test interruptions, verify workflow outcomes, and monitor latency as they shipped voice agents at scale.
My Bottom Line on Vapi AI Pricing
I'd choose Vapi for an engineering-led team that wants runtime control and can handle layered billing. I wouldn't choose it for a team that wants bundled pricing or low operational overhead. The headline rate only makes sense once you model providers, telephony, concurrency, and enterprise needs.
Ready to pressure-test your Vapi stack? If you're building on Vapi, manual calls and thin review coverage won't tell you enough. See how Cekura adds pre-production simulations and production QA on top of a live voice stack.
Frequently Asked Questions
How Much Does Vapi AI Cost?
Vapi's self-serve call pricing starts at $0.05 per minute. Real spend rises once you add STT, LLM, TTS, telephony, and any extra concurrency or enterprise requirements.
Does Vapi Have a Free Plan?
Vapi doesn't offer a permanent free plan on its current public pricing pages. The public setup centers on self-serve and Enterprise, with a separate startup grant for eligible voice AI startups.
What Does Vapi's $0.05 Per Minute Cover?
The $0.05/minute rate covers Vapi's call orchestration layer. It doesn't represent the entire cost of the underlying voice stack.
Does Vapi Offer Enterprise Pricing?
Yes, Vapi offers custom Enterprise pricing. The enterprise motion focuses on scale, reliability, support, and compliance rather than a public per-minute list rate.
Can You Lower Vapi Costs With Your Own Provider Accounts?
Yes, if you validate your own provider keys, Vapi stops billing you for that provider. The provider bills you directly instead.
Is the Vapi Startup Program Still Available?
Vapi still has a public startup grant page, but applications are on hold. The grant advertises up to 90,000 free minutes for eligible startups.
What Is Vapi Best For?
Vapi is best for technical teams building programmable voice products. It fits teams that want control over the voice stack and can manage layered billing.