RexaThe short answer
The practical difference is how the bill is assembled. Vapi charges 5¢ per minute for platform hosting and passes speech-to-text, language model and text-to-speech costs through at cost, so your true per-minute cost depends on which models you pick. Rexa.ai charges 10¢ per minute with those components and recordings already included, plus the phone line on its default setup, where Rexa provisions the number; teams who prefer to keep their own carrier account can bring it, and that carrier bills them directly. Vapi suits teams who want model-level control; Rexa suits teams who want one predictable number.
Last reviewed
Rexa.ai and Vapi at a glance
| Rexa.ai | Vapi | |
|---|---|---|
| Headline rate | 10¢ per minute, all-inclusive | 5¢ per minute platform fee (Build plan) |
| Speech, model and voice costs | Included in the per-minute rate | Passed through to the customer at cost, on top of the platform fee |
| Phone numbers | Included; bring-your-own telecom also supported | Concurrency priced per line (10 call concurrency at 10 per line/month on Build); number provisioning not stated on the pricing page |
| Cost predictability | One rate covers the call | Varies with model selection, since model costs are pass-through |
| Inbound and outbound | Both, on the same platform and the same rate | Calls and SMS/chat channels listed; pricing page does not split inbound from outbound |
| Model choice | Managed stack; the platform selects and runs the models | Developer picks the models and pays their cost directly |
Who Vapi suits better
Vapi is the better fit if you want to choose your own speech and language models per agent, are comfortable reconciling several usage bills, and value model-level control over a single predictable rate.
Who Rexa.ai suits better
Rexa.ai is the better fit if you want one number on the invoice, need phone numbers and live human handoff included rather than assembled, and would rather not tune a model stack to control cost.
Frequently asked questions
- Is Vapi cheaper than Rexa.ai?
- Vapi’s headline rate is lower at 5¢ per minute versus 10¢, but the rates are not measuring the same thing. Vapi’s figure is a platform fee and excludes speech-to-text, language model and text-to-speech, which are passed through at cost. Rexa’s 10¢ includes them, plus the phone line when Rexa provisions the number, which is the default. Which is cheaper depends entirely on the model stack you choose.
- Can I bring my own phone numbers to Rexa.ai?
- Yes. By default Rexa provisions the number and it is included in the per-minute rate, but if you already have a telecom account you can keep it and run calls on your existing numbers. There is no additional platform charge either way; on the bring-your-own path your carrier bills you directly for their per-leg cost.
- Does Vapi support live transfer to a human?
- Vapi’s pricing page does not document human handoff behaviour, so we do not make a claim about it here. Check Vapi’s current documentation for the authoritative answer. Rexa.ai includes live warm transfer and callback continuity in the base rate.
Scope, limitations, and sources
This compares publicly documented pricing and capabilities only. It does not compare latency, voice quality, model accuracy, or reliability — those depend on configuration and region and cannot be established from a pricing page. Vapi figures were read from Vapi’s own pricing page on the access date below and describe the Build plan; Scale-plan and volume terms are negotiated and not public. Vapi’s pricing page does not state whether phone numbers are included, so that row records what the page does and does not say rather than asserting a fact.
Sources
- Vapi Build-plan rate of $0.05/min, per-line concurrency pricing, and the statement that model costs are passed on to the customer. vapi.ai (accessed )
- Rexa.ai per-minute rate and what it includes. rexa.ai (accessed )
Spotted something out of date or wrong? Tell us on the contact page and we will correct it and update the review date.