Vapi alternatives and the real Vapi pricing

On Vapi's own calculator that comes to $82 to $129 for 1,000 minutes a month, or 8 to 13 cents a minute. At 20,000 outbound minutes on Twilio, the bill is about $1,950 to $2,890 a month, and $1,000 of it is Vapi's fee. Every price below comes from each vendor's own pricing page or docs, checked October 23, 2026.

Vapi alternatives at a glance

Vapi alternatives fall into three groups: managed platforms (Retell, Bland, ElevenLabs, Synthflow), cheaper building blocks (Twilio ConversationRelay, Telnyx, Ultravox) and open-source frameworks you host yourself (LiveKit Agents, Pipecat). Here's 5,000 minutes a month on each one's published rate.

The cheap rows charge for less of the stack and hand the rest to you. If you're new to the parts (phone line, speech-to-text, model, voice and the agent layer), our AI voice agent explainer walks through each one.

What is Vapi?

Vapi is an API-first voice AI platform for developers: you pick a speech-to-text engine, an AI model and a voice from its provider lists, and Vapi runs the call, the turn-taking and the tools in the middle. Its FAQ puts typical end-to-end latency around 800 milliseconds and says there's no on-premise option.

Developers like it for choice: 11 speech-to-text providers, 16 model providers and 15 voices, per its provider docs. The same choice makes the bill hard to predict. On Vapi's community forum the most common threads are about cost ("Vapi seems expensive how to justify?") and calls running slower than the dashboard shows. Those are the two reasons people look for alternatives.

Vapi pricing: what a minute really costs

Vapi pricing is a $0.05-a-minute hosting fee plus four provider costs: speech-to-text, the AI model, the voice and the phone line. Vapi bills the providers "at the provider's listed price with no markup," per its pricing page. On its calculator, 1,000 minutes with Deepgram, OpenAI and ElevenLabs costs $82 to $129 a month.

Source: Vapi pricing page, October 23, 2026.

Your model and voice choices move the middle three lines; the hosting fee never moves. You can bring your own API keys for the providers, and Vapi then stops billing for that provider, per its provider keys docs. The hosting fee still applies on top.

Vapi price at 1,000, 5,000 and 20,000 minutes

The Vapi price for a real month runs from about $82 at 1,000 inbound minutes to about $2,890 at 20,000 outbound minutes, once phone minutes and a package are added. The worked bills below use Vapi's calculator ranges, Twilio's US voice rates ($0.0085 in, $0.014 out, $1.15 a number) and the Core package for its 10 concurrent calls.

Two things push real bills above these. Twilio rounds every call up to the next full minute, so lots of short calls cost more than their talk time. And Vapi doesn't publish its own billing increment. Users on Vapi's community forum report paying 19 to 35 cents a minute all-in, and the gap is usually the model, the voice or the carrier.

Outbound AI calls have their own rules: the FCC treats AI voices as an "artificial or prerecorded voice" under the TCPA (FCC 24-17), so you need prior consent. This isn't legal advice; our TCPA rules for sales calls and AI cold calling guides cover it.

Vapi AI pricing plans and add-ons

Vapi AI pricing has one usage rate and three optional "Success Packages" that add concurrency, support and data retention on top. Usage only is $0 with $5 in free credit. Credits are prepaid, and at a $0 balance with auto-reload off, Vapi freezes the account, per its billing docs.

Extra call lines are $10 a month each and extra organizations $20. The free Vapi number is US-only and inbound-only, per Vapi's free telephony docs, so outbound calling means importing a Twilio, Telnyx or other carrier number. The EU region and SSO are Premier only.

Vapi vs Retell

Vapi vs Retell comes down to the model you pick: Retell charges $0.055 a minute plus the model and voice, sells no monthly packages, and includes 20 concurrent calls, while Vapi charges $0.05 plus every provider and includes 4. Retell's pricing calculator opens at $0.15 a minute before telephony; with GPT 5 mini it's about $0.078, level with Vapi's low end.

Pick Vapi if you have a developer and want to choose every provider, bring your own keys and control the call flow in code. Pick Retell if you want an agent live faster, a knowledge base and batch calling built in, and room for 20 calls at once without buying lines. The full Retell bill, add-ons included, is in our Retell AI pricing breakdown.

Bland AI pricing vs Vapi

Bland AI pricing is a single connected rate per plan: $0.14 a minute on the free Start plan and $0.12 on Build at $299 a month, per Bland's billing docs. The docs also list a $499 Scale plan at $0.11, but it no longer appears on Bland's pricing page. The model and voice are included, because Bland runs its own. Telephony is billed separately.

The catch is caps: Start allows 100 calls a day and 10 at once, Build 2,000 a day and 50 at once. Bland also charges a $0.015 minimum per outbound attempt on its lines and $0.03 to $0.05 a minute for transfers. It's easier to forecast than Vapi and fits a high-volume outbound team on Build or above. For a few hundred inbound calls, Vapi or ElevenLabs costs less. Every Bland line item is in our Bland AI pricing guide.

Vapi vs ElevenLabs Agents

Vapi vs ElevenLabs comes down to one flat rate against a stack of parts: ElevenLabs Agents charges $0.08 a minute on every paid plan, covering speech-to-text, voice and the agent, with the AI model passed through at cost and no telephony fee. Plans run from $6 a month (75 minutes) to $990 (12,375 minutes), per its agents pricing page.

On Vapi, hosting plus Deepgram plus an ElevenLabs voice comes to 7.4 to 8.4 cents a minute before the model, so the price is a wash. If you'd use ElevenLabs voices on Vapi anyway, ElevenLabs Agents means one less vendor and 20 to 40 calls at once on Pro to Business. Two catches: calls over your concurrency limit bill at double, $0.16 a minute, and you bring your own carrier.

Synthflow pricing

Synthflow pricing is now enterprise-only: contracts "start at $30,000 annually," and Synthflow publishes no per-minute rate, included minutes or concurrency on its pricing page. Its docs say the older Pro, Growth and Agency plans "are no longer sold." Archived copies of the page show the switch happened between June and July 2026.

Most "Synthflow pricing" articles still quote the old $0.09-a-minute pay-as-you-go plan, which is gone. At $2,500 a month minimum, Synthflow is off the list for a small team testing a voice agent. Agencies also lost in-product rebilling on September 15, 2026.

Twilio ConversationRelay and other per-minute options

Twilio ConversationRelay costs $0.07 a minute on top of regular Twilio voice minutes. Twilio handles speech-to-text and the voice; you bring the AI model and host the server that runs the conversation, per its ConversationRelay docs. It suits a team already on Twilio with a developer and no wish for a middle platform.

Telnyx voice AI

Telnyx voice AI agents cost $0.05 a minute for orchestration, speech-to-text and text-to-speech, plus model tokens and the carrier from $0.0032 a minute, per Telnyx's AI pricing page. Telnyx says a production agent lands near $0.06 a minute. The fine print: the voice engine rounds each call up to 60 seconds, which hurts on short calls. Pay as you go allows 500 calls at once, and US and Canadian numbers port in free from Vapi or Retell.

Ultravox AI

Ultravox AI, now owned by Inworld, charges $0.05 a minute after 30 free minutes, with SIP at half a cent a minute, per its pricing page. Pay as you go is capped at 5 concurrent calls. The $100-a-month Pro plan lifts the cap and never applies surge pricing.

Vapi open source alternatives: LiveKit Agents and Pipecat

The main Vapi open source alternatives are LiveKit Agents and Pipecat, two free frameworks for building the same kind of phone agent in code. You run them on your own servers or their clouds, and pay the speech, model, voice and carrier vendors directly. Vapi itself can't be self-hosted.

LiveKit Agents is Apache 2.0, in Python or Node, and connects to carriers like Twilio and Telnyx over SIP. Pipecat is BSD 2-Clause, maintained by Daily, in Python, with more than 150 AI services plugged in. Both ship turn detection, the hard part of waiting until the caller has finished talking.

LiveKit's prices are from its pricing page, and Pipecat Cloud's from Daily's pricing page.

LiveKit vs Pipecat

LiveKit vs Pipecat is a choice of foundation. LiveKit is a full real-time media stack plus an agent framework, with its own phone numbers, SIP and a cloud that scales to hundreds of calls. Pipecat is a lighter Python pipeline that runs on any transport, LiveKit and Twilio included. Both need an engineer to build and keep running, the cost the per-minute price hides.

Should you rent a Vapi voice agent or own one?

Rent a Vapi voice agent while volume is low and the job is simple; own one when the platform fee becomes a real line on your bill. Vapi's 5-cent hosting fee is 39% to 61% of the cost of every minute on its own calculator. At 20,000 minutes a month it's $1,000 a month, or $12,000 a year, before the AI parts.

Renting is the better call at a few hundred to a few thousand minutes a month, one inbound flow like an AI receptionist, and nobody to maintain software. Vapi on usage only or ElevenLabs Agents does that well. If your AI phone spend is under $1,000 a month, keep renting.

Owning works the other way. The agent itself (conversation logic, scripts, CRM and calendar hookups, recordings and transcripts) runs on your own hosting, built on open frameworks like LiveKit Agents or Pipecat. You still rent a phone carrier and AI models, because every voice agent does, but they're billed as utilities at the vendor's rate with no per-minute platform markup. You can swap any vendor, and the code and call data are yours.

That's what Cleanform builds: an agent you own, on your hosting, wired into the CRM and calendar you already use. It's priced to be typically less than you spend today, paid monthly, and the build payments have an end date. Here's how pricing works. If you're weighing receptionist costs specifically, our AI receptionist cost guide runs the numbers.

Sources

All checked October 23, 2026. Prices change; check the vendor's page before you buy.

• Vapi pricing page and FAQ
• Vapi provider keys, billing and credits, free telephony and voice providers
• Twilio US voice pricing, Twilio billing increments and ConversationRelay docs
• Retell AI pricing
• Bland AI billing docs
• ElevenLabs Agents pricing
• Synthflow pricing
• Telnyx conversational AI pricing
• Ultravox pricing
• LiveKit pricing and LiveKit Agents on GitHub
• Pipecat on GitHub and Pipecat Cloud pricing
• FCC 24-17, AI-generated voices under the TCPA

The SaaS cost calculator adds up your whole software stack in about two minutes. If your phone, AI and CRM bills come to a few thousand a month, book a free audit and we'll show you what an owned system would replace.