Vapi AI Pricing in 2026, Plans and the Real Cost Per Minute

Vapi AI pricing is usage-based. With Vapi’s voice agent pricing, you pay a $0.05-per-minute hosting fee, plus the cost of the speech-to-text, language model, voice and telephony providers each call runs on.

Vapi now also sells new customers optional Success Packages for support and capacity, with uptime commitments on the Pro and Premier packages.

At HappyRobot, we deploy AI agents, which we call AI workers, for more than 150 enterprises. For an enterprise on Vapi, the per-minute rate is only one part of the budget. Concurrency and compliance add-ons are priced on top of it, and the engineering time to keep a call flow stable is a cost of its own.

Let’s break down what Vapi actually costs beyond its per-minute rate. That means counting the engineering hours it takes to turn a voice API into an agent that holds up in production, because that’s an entirely different number that many businesses fail to budget for.

How Vapi's Pricing Model Works

Vapi charges its own per-minute hosting fee for running the call. On top of it, you pay for the speech-to-text, language model and voice providers each call runs on, plus transport for the phone or web connection.

This goes to show you will still be paying a stack of charges rather than a single number.

The Vapi Hosting Fee

Vapi charges a hosting fee of $0.05 per minute. Its pricing page no longer lists a separate SMS or chat rate. Without a package, you get $5 in free credits and 4 concurrent calls. Each extra line costs $10 per month.

The hosting fee covers Vapi’s side of running the call. It keeps the call connected and manages the handoff between the speech and language models in real time. It does not cover the models that run the call, such as speech recognition or language models.

What You Pay on Top of It

Vapi passes model costs through at the provider’s listed price with no markup, on every package. You can also bring your own provider keys.

A call running a premium language model and a premium voice provider costs significantly more per minute than the same call running a cheaper stack, even though Vapi’s own hosting fee stays flat at $0.05 either way.

Moreover, Voice (text-to-speech) is billed by the character of spoken text. Transport, which is the phone or web connection, is a separate line. Calls on Vapi’s own numbers or SIP carry no transport fee, while Twilio adds $0.008 a minute inbound and $0.014 outbound.

Picking providers also means planning for when one slows down or goes offline mid-call. HappyRobot is model-agnostic and fails over automatically across language model, text-to-speech and speech-to-text providers when one has an outage or its latency degrades, so the provider stack is ours to manage, not yours.

Success Packages for Support, SLAs and Capacity

A Success Package doesn’t change how Vapi bills usage. It sits on top of usage and adds support, service-level agreements, more concurrent calls and compliance controls. Core is $29 a month. Pro is 10% of your Vapi hosting fees, with a $999 monthly minimum. Premier is custom and sold directly by Vapi’s sales team.

Success Packages are only available to new Vapi customers. Customers who signed up before the change keep their existing billing structure and terms.

How Much Does Vapi AI Cost Per Minute?

Vapi charges a flat $0.05 hosting fee per minute. With the default providers in Vapi’s pricing calculator (Deepgram, OpenAI and ElevenLabs), a minute costs about $0.08 to $0.13 in total before transport.

The figures above reflects the estimates for a typical minute, but a real call can cost more.

Vapi's own documentation says long calls usually cost more than the estimate, because the growing conversation history is sent to the model again with every request. Longer prompts and more tools raise the model cost too. On the cheapest default stack, the hosting fee is about 61% of the per-minute cost, so the model and voice choices move the total.

The $0.05 hosting fee stays the same on every call, so the language model and voice you pick decide whether a minute costs closer to $0.08 or $0.13. Speech-to-text changes the total the least. Deepgram, the default provider, costs $0.0095 to $0.0099 a minute through Vapi, which you can compare with Deepgram’s own pricing.

One Thing the Per-Minute Price Doesn’t Include: Build Time

A developer on Reddit documented 760 calls and over 100 hours of work building a Vapi agent. Their finding: the AI itself was roughly 20% of the effort. The rest went into error handling and the automation workflows needed to feed call outcomes back into a CRM.

It speaks of building an after-hours agent to help customers look up specific items in a Shopify store with 4,500+ products. It needed a separate workflow to pull the right products from a disorganized catalog and hand the AI a clean answer, all within 800 milliseconds, before the call went silent. That one piece took days to get stable.

The replies say the same thing. A developer switched platforms because when a Vapi call broke, the logs didn’t show what went wrong. Another posted a live bug: a workflow that worked on a test URL stopped firing the moment it went to production, with no logs to explain why.

This kind of build time isn’t unique to Vapi. It shows up across modular voice platforms, which is why it belongs in the budget before you compare Vapi’s advertised rate against a fully managed alternative.

Visibility is the part we built for. On HappyRobot, every run keeps its transcript and recording, northstars score sampled live sessions against pass/fail criteria, and once a production failure is fixed, that session is locked in as a regression test.

Book a demo with HappyRobot and bring the call flow that took your team the longest to get stable. We’ll show you an AI worker running it, from the first ring to the update in your system.

How Much Does Vapi Cost Per Month?

To estimate a monthly bill, multiply the per-minute cost by your monthly minutes, then add the package fee. The table uses Vapi’s default calculator stack (Deepgram, OpenAI and ElevenLabs). It assumes calls run on Vapi’s own phone numbers, which carry no transport fee, and it excludes add-ons.

If the calls touch health data, add $2,000 a month per organization for Vapi’s HIPAA add-on, which comes to $24,000 a year. Each concurrent line beyond your package adds $10 a month. Pro’s 10% fee stays at its $999 minimum until you pass about 200,000 minutes a month.

Are you in the Pro rows?
Talk to our team before you commit to the $999 minimum. We’ll price the same call volume with AI workers our engineers build and run for you.

Are you in the Pro rows? At that volume, every concurrent line past 30 costs another $10 a month. On HappyRobot, inbound calls are processed immediately with no concurrency limits, and DHL already runs millions of voice minutes a year through our AI workers.

Talk to our team before you commit to the $999 minimum. We’ll price the same call volume with AI workers our engineers build and run for you.

What Are the Pricing Plans for Vapi?

Vapi has one usage-based pricing plan and three optional Success Packages that sit on top of it: Core, Pro and Premier. That gives four ways to buy Vapi: usage only, or usage with Core, Pro or Premier. Usage is billed the same way on every option, so the package changes support, capacity and compliance rather than the per-minute rate.

Pricing Plans for Vapi

Is Vapi AI Free or Paid?

Vapi is paid, but you can start without a subscription. New accounts get $5 in free credits, and once those run out, every minute is billed at the usage rates above. Without a package, an account includes 2 organizations, 1 Vapi phone number, 4 concurrent calls and 14 days of raw data retention.

Support comes through the Discord community and email, with no response-time commitment. Zero data retention isn’t included until you add a package.

On HappyRobot, the OpenAI and Anthropic models our AI workers run on are used under zero data retention, and your data is never used to train models.

Core, Pro and Premier Success Packages

Model costs pass through at cost on all three packages. Pro’s percentage is calculated on hosting fees only, not on model costs.

Core is a flat $29 a month. Pro starts at $999 a month, and that buys 30 concurrent calls, a 99% uptime SLA and basic roles. Premier is the only package with SSO, custom limits and a 99.9% uptime SLA.

Vapi Add-Ons for HIPAA, Concurrency and Organizations

Add-ons attach to any package, including usage only. Vapi’s HIPAA add-on costs $2,000 per organization per month. HIPAA mode is an organization-wide setting, so you can’t switch it on for a single agent or phone number.

Vapi HIPAA Add-On

Once on, it applies to every assistant in that organization and changes how Vapi handles call logs, recordings and transcripts. It also limits the organization to HIPAA-compatible model, voice and transcriber providers. Vapi includes a standard DPA and emails a Business Associate Agreement (BAA) for signature when you enable the add-on. Extra concurrent lines cost $10 each per month, and extra organizations cost $20 each per month.

Vapi states it is SOC 2 Type II certified and GDPR compliant. PCI compliance is set per assistant instead. You turn on the PCI Compliance toggle in each assistant’s settings, and that assistant is then limited to PCI-compliant model, voice and transcriber options.

If your security review will be harder than the pricing conversation, HappyRobot's trust center lists SOC 2 Type II, ISO 27001, GDPR and HIPAA attestations, and alignment with the EU AI Act, NIST CSF and DORA. PCI DSS 4.0.1, CCPA and CSA STAR are also listed, along with an AI Red Team & Pentest 2026 report.

Is Vapi Worth It, or Should You Deploy a Voice AI Agent Instead?

Vapi is worth its modular pricing if your team has the engineers to assemble, test, and maintain a voice pipeline. If you don’t have those engineers, a low per-minute rate buys you very little.

VAPI average cost per call | Source: YouTube

You are still paying for the build, just through engineering salaries rather than usage fees. For teams without those engineers, we compare the other Vapi alternatives in a separate guide.

Vapi fits teams that want to choose their own speech, language, and voice providers and are ready to own error handling and the automation that turns a finished call into a CRM update or a follow-up message. That’s a real control, but it comes at the cost of the engineering time required to build and maintain it.

Vapi builds for engineering teams, and its homepage headline says: “Voice agents for builders.” One of its customer stories is headlined around engineering headcount: 4 engineers ramped Spring Venture Group to the volume of 50 call center agents in one week. The two customers it quotes by name on its homepage are Ring’s VP of Software Development and Kavak’s Chief Product & AI Officer. Vapi also runs a healthcare page for patient scheduling, triage and benefits verification.

Voice agents for builders

We build for the people accountable for the operation itself. That’s the COO or CFO who owns the labor line, and the VP who runs operations, sales, customer care or collections and is measured on whether the work gets done. Our AI workers take the calls that keep an operation running, such as technician dispatch and appointment scheduling, and complete the task in the systems behind them.

Why Is Vapi So Expensive?

Vapi gets expensive through the charges that grow with use, while its $0.05 hosting fee stays flat. The model, voice and transcription charges rise with every minute, and a long call costs more per minute than the estimate. Extra concurrent lines and HIPAA are separate line items, and longer data retention only comes with a higher package. Your own team’s engineering time to build, test and maintain each call flow adds to that, and it doesn’t appear on a Vapi invoice.

In a HappyRobot deployment, our Forward Deployed Engineers build, test and deploy that work on-site with your team. At WWEX, all 6 primary agents were in production 6 weeks after kickoff.

HappyRobot as a Vapi Alternative for Enterprise Operations

HappyRobot is an AI agent platform for enterprise operations. Our Forward Deployed Engineers build and deploy AI workers with your team, so you don’t assemble or operate the stack yourself. Our voice AI workers speak 30+ languages, and HappyRobot is sold on custom enterprise pricing.

At Naturgy, the Spanish utility, our AI worker is built to answer 100% of incoming calls at the same time with no wait. It scored 9.4 out of 10 for customer satisfaction in its first weeks and went from proof of concept to production in three months.

At Encompass, a home repair parts company, our AI workers handle about 1,700 LG-related calls a week and resolve 60 to 70% of them without a person, with 93% non-negative customer sentiment.

In logistics, our AI workers at Kuehne+Nagel handle 78% of connected calls end to end, and the team’s capacity has grown by 47%.

When a call needs something changed, the AI worker makes the change itself, booking or updating records in the systems you already run. When a call needs a person, it does a warm transfer and passes along a summary of the conversation.

Vapi figures come from Vapi’s pricing page and documentation as of October 7, 2026. Confirm current rates on the Vapi pricing page before you budget. HappyRobot figures come from its published customer stories.

For a full side-by-side on architecture and workflow depth, see Vapi vs. HappyRobot.

Book a demo with HappyRobot to see an AI worker handle your highest-volume call end to end, including what it updates after the caller hangs up.

The Bottom Line on Vapi Pricing

Vapi publishes its hosting fee, and the rest of the bill depends on the stack you choose. On the default stack, 1,000 minutes a month costs $82 to $129. At 300,000 minutes, it’s $26,040 to $40,170 a month before HIPAA or extra concurrent lines, and your own engineers build and maintain every call flow.

That model suits a team that wants to build and own every layer. If you need the work done and the calls answered, book a demo with HappyRobot and we’ll scope your first AI worker with you.