Vapi's headline price in 2026 is $0.05 per minute for platform hosting, but that covers only orchestration — transcription, the language model, voice synthesis, and telephony are billed separately at provider cost, which pushes a realistic 1,000-minute stack to roughly $82–$129 per month on Vapi's own calculator. Structured plans start at $29/month (Core) with a $999/month minimum for Pro, and HIPAA-eligible data handling adds $2,000/month. If you want a done-for-you voice agent rather than a build-it-yourself API platform, Agents by AIQ builds and operates AI receptionists and phone agents for small businesses on a month-to-month basis.
Bottom line: Vapi is priced as developer infrastructure: cheap at the headline rate, but the real bill assembles from five sources (hosting, transcription, LLM, voice, telephony) and requires engineering time to manage. For small businesses without developers, a managed alternative usually delivers a working phone agent faster and with a predictable bill.
TL;DR
- Vapi's published platform fee is $0.05 per minute of call time, billed to the second, with $5 in free starter credits.
- That fee covers hosting/orchestration only — transcription, the LLM, voice synthesis, and telephony are passed through at provider cost.
- Vapi's own calculator estimates a 1,000-minute month at $82–$129/mo all-in (Deepgram + OpenAI + ElevenLabs + free SIP transport).
- Structured plans: Core at $29/month, Pro at a $999/month minimum, Premier by custom quote.
- HIPAA-eligible data handling is a $2,000/month add-on; extra concurrency is $10/line/month and extra organizations $20/month.
- Independent testing puts the real all-in runtime cost between roughly $0.07 and over $1.00 per minute depending on the model stack.
- Build time is a hidden cost: one developer documented 760 calls and 100+ hours of work to get a single agent production-ready.
Head-to-head
| Dimension | Vapi | Agents by AIQ |
|---|---|---|
| Pricing | $0.05/min hosting + provider pass-throughs; Core $29/mo; Pro $999/mo min; HIPAA +$2,000/mo | Month-to-month engagement scoped per business |
| Ease of use | Developer-first; light visual builder, most power in code; reviewers call the dashboard overwhelming | Done-for-you — the AIQ team builds and operates the agentwinner |
| Features | Full model catalog, testing/simulation, monitoring, endpointing, webhooks, custom tools, unlimited agentswinner | AI receptionist, voice agents, sales/SDR, support, email, marketing agents, appointment setters, workflow automation |
| Support | Discord + email (no SLA) on free tier; 2-business-day SLA on Core; 1-business-day + Slack on Pro; named team on Premier | Operated by the AIQ Labs teamwinner |
| Integrations | API-native: BYO SIP (Plivo, Telnyx, Zadarma), Twilio/Vonage/Telnyx telephony, webhooks, custom endpointswinner | Integrated with the tools the business already uses |
| Setup | Self-build: one developer documented 760 calls and 100+ hours for one production agent | AIQ designs, builds, connects, and runs the agent for youwinner |
| Compliance | HIPAA-eligible handling $2,000/mo; ZDR $1,000/mo (reported); SSO/RBAC on Pro and above | Handled as part of the managed build |
Scores
Vapi
vapi.ai ↗Vapi is a developer-centric voice AI platform for building, testing, and deploying voice agents into apps, CRMs, and phone systems. Every capability is accessible through its API — hence the name — and the platform's core pitch is 'bring your own stack': you choose the speech-to-text engine (Deepgram, AssemblyAI, Whisper), the language model (OpenAI, Anthropic, Google), the voice provider (ElevenLabs, Azure, Play.ht), and the SIP/telephony provider (Twilio, Telnyx, Vonage, Plivo), while Vapi orchestrates the real-time listen-think-speak loop. Typical setups run 500–800ms of response latency, and Vapi advertises sub-500ms average latency across 1M+ calls per month at enterprise scale. The platform includes endpointing, interrupt detection, backchanneling, call testing and simulation, and critical-issue monitoring. It is genuinely powerful in the hands of an engineering team, but it is priced and designed as infrastructure, not as a finished product: there are no seats or feature tiers, the bill assembles from five separate provider sources, and non-technical teams routinely find the dashboard overwhelming.
| Plan | Price | What you get |
|---|---|---|
| No Success Package (usage-based) | $0/mo | $0.05/min hosting, $5 free credits, 2 organizations, 1 phone number, 4 concurrent calls, 14-day raw data retention, Discord + email support (no SLA) |
| Core | $29/mo | 2 organizations, 5 phone numbers, 10 concurrent calls, 30-day retention, ZDR included, 2 business day email SLA |
| Pro | $999/mo | 10% of hosting fee, $999/mo minimum, 10 organizations, 10 phone numbers, 30 concurrent calls, 180-day retention, RBAC + ZDR, Slack support, 99% uptime SLA |
| Premier | $0/mo | Custom quote; named support team, 1-hour P0 SLA, 99.9% uptime, custom RBAC + ZDR + SSO, dedicated SIP |
| HIPAA add-on | $2000/mo | HIPAA-eligible data handling, standard DPA and BAA |
| Additional concurrency | $10/mo | $10 per extra concurrent line per month |
| Additional organization | $20/mo | $20 per extra workspace per month |
Pros
- Very low entry barrier: $5 free credits and pay-only-for-minutes usage billing with no required base subscription
- Unmatched model flexibility — mix and match STT, LLM, TTS, and SIP providers, or bring your own API keys at zero markup
- Strong real-time performance: sub-500ms average latency, endpointing, interrupt detection, and backchanneling
- Unlimited agents on every plan; you own and operate your agents, models, and data
- Enterprise-grade options including 99.9% uptime SLA, 1-hour P0 response, SSO/RBAC/ZDR, and dedicated SIP on Premier
- Telephony transport via Vapi SIP is free; Twilio inbound is $0.008/min and Telnyx $0.0055/min
Cons
- The $0.05/min headline rate excludes transcription, LLM, voice, and telephony — real runtime cost runs roughly $0.07 to over $1.00 per minute depending on the stack
- HIPAA-eligible handling costs $2,000/month, which alone can exceed call costs for a regulated small business
- Steep for non-developers: reviewers call the dashboard overwhelming, and one developer documented 100+ hours of build work for a single production agent
- Concurrency is limited on lower tiers (4 concurrent calls on the free tier, 10 on Core) with $10/month per extra line
- No native SSO, RBAC, or SOC 2 on the usage-based tier; those require the $999/month Pro minimum or Scale contract
Agents by AIQ
agentsbyaiq.com ↗Agents by AIQ is the agent-focused arm of AIQ Labs, offering done-for-you AI agents for small and mid-size businesses. Rather than handing you a toolkit, the team designs, builds, connects, and operates the agent for you: AI receptionists and phone answering on a real phone number, browser-based voice agents, sales follow-up and SDR agents, customer support agents, email agents, marketing agents, appointment setters, and workflow automation — integrated with the tools the business already uses. The positioning is explicitly not another AI wrapper or DIY toolkit: it is a managed build run by the team behind the AI Business Sites platform, aimed at owner-operators and small teams in trades, legal, healthcare, insurance, real estate, ecommerce, and professional services who are losing missed calls, slow lead follow-up, and manual busywork. Engagements are month-to-month, the client owns everything, and the starting point is a scoping call to define the agent for your business.
Pros
- Done-for-you: AIQ Labs designs, builds, connects, and runs the agent — no DIY provider wiring or prompt engineering
- Month-to-month terms with no long-term contract; the client owns everything
- Covers the full small-business agent surface: phone answering, sales follow-up, support, email, marketing, appointments, and automation
- Integrates with the tools the business already uses
- Backed by the team behind the AI Business Sites platform
Cons
- Not a self-serve platform — you work with the AIQ team rather than building agents yourself
- Scoped per business, so it is not a plug-and-play catalog of templates
- Best suited to owner-operators and small teams rather than developer teams wanting raw API control
Pricing compared: what $0.05/minute actually buys
Hidden costs: build time and concurrency
Which model fits which business
Which should you pick?
FAQs
How much does Vapi cost per minute in 2026?
Vapi charges $0.05 per minute as its platform hosting fee, prorated to the second. That fee covers only orchestration — transcription, the language model, voice synthesis, and telephony are billed separately at provider cost with no Vapi markup. Vapi's own calculator estimates a 1,000-minute month at $82–$129 total using Deepgram, OpenAI, and ElevenLabs, and independent testing puts the realistic all-in range between roughly $0.07 and just over $1.00 per minute depending on the stack you choose.
Does Vapi have a free plan or free trial?
Vapi does not have an ongoing free plan. New accounts on the usage-based tier get $5 in free credits to get started, with no required base subscription — you pay only for the minutes that run. The free tier includes 2 organizations, 1 included Vapi phone number, 4 concurrent calls, 14-day raw data retention, the full model catalog, and support via the Discord community plus email without an SLA.
What are Vapi's paid plans?
Vapi offers three structured Success Packages on top of its usage-based tier. Core costs $29/month and includes 5 phone numbers, 10 concurrent calls, 30-day retention, ZDR, and a 2-business-day email SLA. Pro runs at 10% of the Vapi hosting fee with a $999/month minimum, adding 30 concurrent calls, 180-day retention, RBAC + ZDR, dedicated Slack support, a 1-business-day SLA, and a 99% uptime SLA. Premier is custom-quoted with a named support team, 1-hour P0 response, and 99.9% uptime.
Why is Vapi's real bill higher than $0.05/minute?
Because the $0.05 is only Vapi's hosting and orchestration fee. You separately pay for speech-to-text (e.g. Deepgram at $0.0095–$0.0099/min), the intelligence model (OpenAI at $0.0077–$0.0452/min), voice synthesis (ElevenLabs at $0.0146–$0.0238/min), and telephony (Twilio inbound $0.008/min, outbound $0.014/min, or free Vapi SIP transport). A premium LLM plus a premium voice can push the all-in cost past $1.00 per minute, and add-ons like HIPAA handling ($2,000/month) and extra concurrency ($10/line/month) stack on top.
Is Vapi a good fit for small businesses without developers?
Generally no. Vapi is a developer-first platform where most power lives in code, and reviewers describe the dashboard as overwhelming and have asked for a basic mode. One developer documented over 100 hours of work and 760 calls to get a single production agent running, with most effort going to error handling and CRM workflows. Small businesses without engineering bandwidth typically get a working phone agent faster from a done-for-you service like Agents by AIQ, which builds and operates the agent for you on month-to-month terms.
Sources
- Vapi Pricing — Build Advanced Voice AI Agents
- Vapi — Voice Agents for Builders
- Vapi AI Pricing 2026: The Real Cost Per Minute
- Vapi AI Review 2026 (Retell AI)
- Vapi AI Review in 2026: Pricing, Pros, & Cons (Lindy)
- Vapi Pricing 2026: What It Costs & How to Think About It
- Vapi AI Pricing: What the Per-Minute Model Actually Costs
- 5 Best AI Voice Agents in 2026
- Agents by AIQ