VAPI pricing: per-minute rates, tiers, and fees

VAPI pricing: per-minute rates, tiers, and fees

Voice AI is driving real business results. According to a 2025 report from CMX Today, 97% of SMBs report higher revenue and better client engagement using AI voice agents, and properties using AI cold calling see a 22% sales boost. But to capture that ROI, you need to understand what you're actually paying.

Vapi has emerged as a popular voice AI platform, but its pricing structure requires careful analysis. This guide breaks down Vapi's per-minute rates, tiers, and fees so you can model costs accurately and compare vendors with confidence.

How Vapi pricing works

Vapi uses a tiered subscription model combined with per-minute usage fees. The platform charges a flat rate of $0.05 per minute on its Ad-Hoc plan, but that's just the platform fee. Total costs stack up quickly when you factor in telephony, speech-to-text, text-to-speech, and LLM inference.

According to Dograh AI, Vapi's layered pricing typically lands between $0.23 and $0.33 per minute when all components are included. Synthflow AI reports the effective Ad-Hoc rate at approximately $0.18 per minute.

Vapi pricing tiers at a glance

Sources: Synthflow AI, G2

What's actually driving your costs

Voice AI pricing isn't just about platform fees. Every call touches multiple billable components: PSTN termination to connect calls to real phone numbers, STT to transcribe caller speech, TTS to generate agent responses, and LLM inference to power the conversation itself.

AI voice agent costs can range from a few cents per minute to several hundred dollars per month depending on how these pieces are sourced. Platforms that rely on third-party providers for telephony or speech services pass along markups at each layer. Platforms that own their infrastructure can offer more predictable, transparent pricing.

For context, Retell AI prices voice agents at $0.07+ per minute, while Telnyx offers Voice AI at $0.08 per minute including STT, TTS, and open-source AI on a single platform.

Why full-stack platforms lower total cost

Stitching together separate vendors for telephony, transcription, synthesis, and inference creates complexity and cost. Each integration point adds latency, and each vendor adds margin.

Platforms like Telnyx that unify global communications infrastructure with AI eliminate these inefficiencies. As a licensed telecom provider in 30+ markets, Telnyx connects Voice AI directly to the PSTN without third-party intermediaries. This architectural advantage means fewer markups, lower latency, and simpler cost modeling.

The difference matters at scale. Running AI voice agents at $0.06 per minute translates to roughly $3.60 per hour, making advanced Voice AI accessible for high-volume use cases that were previously limited by cost. Compare that to Vapi's effective rates of $0.18–$0.33 per minute, and the savings compound quickly.

Choosing the right platform

Vapi works for teams exploring voice AI, but scaling deployments often expose limitations. The platform's 2.6 Trustpilot rating reflects mixed experiences, particularly around pricing transparency and support.

When evaluating alternatives, consider whether the platform owns its telephony infrastructure, offers transparent per-component pricing, and can scale without surprise fees. Telnyx's side-by-side comparison with ElevenLabs and Vapi breaks down performance, latency, and cost across these dimensions.

For teams ready to move beyond experimentation, Telnyx offers a full-stack Voice AI platform with built-in telecom, colocated GPU infrastructure for low-latency inference, and usage-based pricing that scales predictably.

Build voice AI that scales

Telnyx is the only platform that unifies carrier-grade communications infrastructure with fully-featured AI. Provision numbers, connect to the PSTN, and deploy Voice AI agents from a single platform at $0.08 per minute.

Explore Telnyx Voice AI pricing or contact our team to model costs for your use case.