Vapi's pricing looks simple — $0.05 per minute — and that number is misleading on its own. It only covers Vapi's orchestration. The models and telephony that actually make a voice agent work are billed separately and stack on top, so the real all-in cost usually lands between $0.07 and $0.25+ per minute. Here's exactly how the bill is built.
The quick answer
Vapi charges a $0.05/min platform fee for orchestration. On top of that you pay for four external components — speech-to-text, the LLM, text-to-speech and telephony. Add them up and a typical agent costs $0.07–$0.25+/min, with premium LLMs and voices pushing it toward $0.30. There's no ongoing free plan (just trial credits), and pay-as-you-go caps you at 10 concurrent calls.
The five cost layers
Vapi is a builder platform, so you assemble the pipeline and pay for each piece. This is the whole reason the sticker price understates the bill.
| Component | Typical cost/min | What it does |
|---|---|---|
| Platform fee | $0.05 | Vapi's orchestration — turn-taking, routing, call logic. |
| Speech-to-text (STT) | ~$0.01 | Transcribes the caller's speech (e.g. Deepgram). |
| LLM | $0.02–$0.20 | The model that reasons and replies — the biggest swing factor. |
| Text-to-speech (TTS) | ~$0.04 | Speaks the reply (e.g. Cartesia, ElevenLabs). |
| Telephony | ~$0.01 | Carries the call over the phone network. |
The LLM is where your cost lives or dies: a cheap fast model keeps you near the floor, while a frontier model can multiply the per-minute rate. Your TTS and STT choices move it too.
Real per-minute math
Add the floor: $0.05 + $0.01 + $0.02 + $0.04 + $0.01 ≈ $0.13/min with budget models. Swap in a premium LLM at $0.20/min and a top-tier voice and you're at $0.30+/min. Most real deployments settle between $0.07 and $0.25 once tuned.
At scale that compounds fast. A moderate enterprise deployment commonly runs $3,000–$6,000/month before setup and maintenance. The lever you control is model choice — picking a lean LLM and an efficient voice is the single biggest cost saving.
Concurrency and HIPAA
Two things catch teams out. First, concurrency: pay-as-you-go limits you to 10 simultaneous calls, and unlimited concurrency requires an Enterprise contract with custom pricing — a real constraint if you expect call spikes. Second, HIPAA: compliance is an add-on of about $1,000/month, so a healthcare agent with moderate usage can total around $1,300/month once you fold in per-minute costs.
Is Vapi worth it?
For developers who want a flexible, model-agnostic platform to build custom voice agents — mixing their preferred STT, LLM and TTS with strong orchestration and integrations — Vapi is genuinely capable and popular for good reason. See the full Vapi review for the feature story.
The honest caveat is forecasting: the stacked, usage-based model makes costs hard to predict until you've measured real calls. Model your per-minute cost with your actual model choices before you scale, and compare against a more managed platform in Vapi vs ElevenLabs or the wider best AI voice generators guide.
Frequently Asked Questions
How much does Vapi cost in 2026?
Vapi's platform fee is $0.05 per minute, but that only covers orchestration. Once you add speech-to-text, the LLM, text-to-speech and telephony, real all-in cost typically lands between $0.07 and $0.25+ per minute depending on the models and voices you choose.
Why is Vapi's real cost higher than $0.05/min?
Because $0.05/min is only Vapi's orchestration layer. You separately pay STT (~$0.01/min), the LLM ($0.02–$0.20/min), TTS (~$0.04/min) and telephony (~$0.01/min). Those stack on top, so the total is the sum of all five components.
Does Vapi have a free tier?
Vapi offers free trial credits to get started, but there's no ongoing free plan for production. Pay-as-you-go begins at the $0.05/min platform fee plus the underlying model and telephony costs, and it caps you at 10 concurrent calls.
How much does HIPAA compliance cost on Vapi?
HIPAA compliance is an add-on of roughly $1,000/month. For a healthcare deployment with moderate usage, total monthly spend can reach around $1,300 once you include per-minute costs and the compliance fee.
Is Vapi worth it?
For developers who want a flexible, model-agnostic platform to build custom voice agents, yes — Vapi's orchestration and integrations are strong. The catch is that its stacked, usage-based pricing is hard to forecast, so model your real per-minute cost before scaling.
How do I lower my Vapi bill?
Model choice is the biggest lever. Use a lean, fast LLM and an efficient voice like Cartesia to stay near the floor, keep responses concise to reduce TTS characters, and only pay for HIPAA or unlimited concurrency if you truly need them.