Skip to content
New: a sample workflow audit report
Scenox Studios
Book a call

Vapi vs Retell vs Bland vs a custom voice AI build

Pick Vapi if your engineers want to choose and swap each model and provider. Pick Retell for a managed stack with a published per-minute rate to budget against. Pick Bland if you want the model and voices inside one rate and an on-prem or VPC option at enterprise level. Build custom when voice is your product or a data-residency rule excludes hosted platforms.

Book a strategy callJump to the table
Last updated Vendor facts as of October 2026

Facts below come from each vendor's own pages, checked October 16, 2026. The pick-by-situation lines are our reading of them.

Voice AI platforms side by side, as of October 2026

Prices are US list prices shown on each vendor's page on October 16, 2026. Plans, rates and limits change often, so confirm on the vendor's site before you commit. Compliance rows are the vendor's own statements, not our audit.

  • What it is

    Sources29131516
    Vapi
    A developer platform for voice agents. Each assistant combines speech-to-text, a language model and text-to-speech.
    Retell
    A platform to build, test, deploy and monitor voice and chat agents. It runs the telephony, speech and model layers for you.
    Bland
    A voice AI platform with agent building, monitoring and enterprise deployment options.
    Custom build
    Open-source frameworks such as LiveKit Agents (Apache-2.0) or Pipecat (BSD-2-Clause), run and hosted by you.

  • Pricing model

    Sources1812
    Vapi
    A hosting fee of $0.05 a minute on pay-as-you-go. Speech, model and voice providers are billed at cost with no markup. Paid packages start at $29 a month.
    Retell
    Pay as you go, per minute. Voice infrastructure is $0.055 a minute, with model and voice priced on top. Headline range $0.07 to $0.31 a minute, with $10 in free credits.
    Bland
    Start: $0.14 a minute of talk time, no platform fee. Build: $0.12 a minute plus $299 a month. Model, speech-to-text and premium voices are inside the rate.
    Custom build
    You pay each provider directly for telephony, speech, model, voice and hosting, plus the engineering time to wire them together.

  • Telephony

    Sources1489121416
    Vapi
    Vapi telephony and SIP, or import numbers from Twilio, Telnyx or Vonage, or connect your own SIP trunk. Carrier fees are billed by the carrier.
    Retell
    Buy a US or Canada number from Retell, or connect your own provider over SIP trunking. Retell lists US Twilio telephony at $0.015 a minute and no charge for custom telephony.
    Bland
    Built-in Twilio at pass-through cost, or bring your own Twilio account or SIP provider. Customers who bring their own telephony pay no transfer fees.
    Custom build
    A SIP trunk, or Twilio, Telnyx, Vonage and others through the framework's transport or serializer layer.

  • Models and voices

    Sources2381012
    Vapi
    Lists dozens of providers, including OpenAI, Anthropic, Google, Deepgram and ElevenLabs. A custom LLM endpoint is supported.
    Retell
    Model and voice choices priced per minute, including GPT, Claude and Gemini models. A custom LLM over WebSocket is supported, and its docs say it gives up some features such as the LLM playground and agent transfer.
    Bland
    The language model and premium voices are included in the per-minute rate. The pages we read do not say which models run.
    Custom build
    Any provider you can call, including open-weight models that you host.

  • Concurrency limits

    Sources1812
    Vapi
    4 concurrent calls on pay-as-you-go, 10 on Core, 30 on Pro. Extra lines cost $10 a line a month.
    Retell
    20 concurrent calls included, then $8 per extra concurrent call a month. Enterprise starts at 50 or more.
    Bland
    Start: 10 concurrent calls and 100 calls a day. Build: 50 concurrent and 2,000 a day. Enterprise is custom.
    Custom build
    Set by your own infrastructure and your providers' limits.

  • Compliance claims (vendor statements)

    Sources15678912
    Vapi
    Trust center lists SOC 2 Type II, SOC 3, GDPR, CCPA, HIPAA and PCI DSS v4.0.1. HIPAA needs a signed BAA and a $2,000 a month add-on.
    Retell
    Docs say SOC 2 Type 1 and Type 2, GDPR, and HIPAA with a signed BAA. The pricing page lists the BAA as an enterprise feature.
    Bland
    Pricing FAQ says SOC 2 Type I and II, GDPR, PCI DSS, and HIPAA-eligible with a signed BAA. The BAA and data residency are listed on Enterprise.
    Custom build
    Nothing is inherited. You certify your own stack, or choose providers that already hold the reports you need.

  • Self-hosting

    Sources15812
    Vapi
    No. Vapi's FAQ says it does not support on-premise deployments.
    Retell
    Not mentioned on the pages we read. A dedicated stable server is listed as an enterprise add-on.
    Bland
    Enterprise lists on-prem and VPC deployment as available.
    Custom build
    Yes. It runs wherever you deploy it.

  • Best fit

    Vapi
    Engineering teams that want to choose and swap each component.
    Retell
    Teams that want a managed stack and one rate to budget against.
    Bland
    Buyers who want one rate and an enterprise deployment option.
    Custom build
    Products where voice is the core, strict residency rules, or volumes where platform fees outweigh the cost of replacing them.

  • What none of them builds for you

    Vapi
    The call flow, the tools that reach your CRM or calendar, testing on real calls and the handoff to a person.
    Retell
    The call flow, the tools that reach your CRM or calendar, testing on real calls and the handoff to a person.
    Bland
    The call flow, the tools that reach your CRM or calendar, testing on real calls and the handoff to a person.
    Custom build
    The same list, plus the infrastructure, monitoring and on-call duty.

Which platform fits which team

Limits of this comparison

  • This table compares platforms. It does not compare call quality. Latency, interruption handling and voice naturalness need a test call with your own script, and we have not benchmarked these vendors against each other.
  • Per-minute rates are not a total cost. Telephony, add-ons, the build and the calls that go to a person all add to it. Our voice agent cost guide shows how to work out cost per call.
  • Compliance rows are the vendors' own statements. A SOC 2 report or a BAA does not make your agent compliant. Call recording and disclosure rules are your responsibility in each country you call.
  • Details such as Retell's per-minute add-ons and Vapi's package limits are easy to misread. Read the live pricing page and ask each vendor to confirm in writing.
  • Plans change. This page reflects what was published on October 16, 2026.

Where Scenox Studios fits

A platform gives you the phone layer. It does not write the call flow, connect it to your calendar or CRM, test it on real calls or decide when a person takes over. We build that layer, on whichever stack suits your call volume and rules, and we say so if a platform alone is enough.

Sources

Each link opens the vendor's own page. We read them on October 16, 2026, and the table cites them by number.

  1. 1
    Vapi pricing

    Hosting fee, packages, concurrency, HIPAA add-on, provider rates

  2. 2
    Vapi documentation overview

    Provider choice per component

  3. 3
    Vapi custom LLM guide
  4. 4
    Vapi SIP trunking guide

    Bring your own SIP trunk

  5. 5
    Vapi FAQ

    Compliance statements and the on-premise answer

  6. 6
    Vapi HIPAA documentation
  7. 7
    Vapi trust center

    Listed frameworks and SOC 2 Type II and SOC 3 coverage period

  8. 8
    Retell pricing

    Per-minute rates, add-ons, concurrency, enterprise features

  9. 9
    Retell documentation introduction

    Telephony options, agent types, compliance statements

  10. 10
    Retell custom LLM overview
  11. 11
    Retell home page

    Compliance and SIP statements

  12. 12
    Bland pricing

    Plans, per-minute rates, limits, enterprise features, FAQ

  13. 13
    Bland documentation
  14. 14
    Bland bring-your-own-Twilio guide
  15. 15
    LiveKit Agents repository

    Apache-2.0 license

  16. 16
    Pipecat repository

    BSD-2-Clause license, listed telephony serializers and providers

Questions people ask

What is the difference between Vapi and Retell?

Vapi bills a hosting fee ($0.05 a minute on pay-as-you-go) and passes speech, model and voice providers through at cost, so you assemble the stack. Retell sells voice infrastructure at $0.055 a minute with model and voice priced on top, and runs the telephony, speech and model layers for you. Both are managed platforms. Prices are from their pages as of October 2026.

Retell vs Bland: which is cheaper?

It depends on your configuration. Bland lists $0.12 to $0.14 a minute with the model and premium voices included, plus telephony. Retell lists a $0.07 to $0.31 a minute range that moves with the model and voice you pick. Price a real call on each, including telephony and add-ons, before you compare. Rates are as of October 2026.

Which voice AI platform can be self-hosted?

From the pages we read, Bland lists on-prem and VPC deployment as available on Enterprise. Vapi's FAQ says it does not support on-premise deployments. Retell's pages do not mention self-hosting and list a dedicated stable server as an enterprise add-on. Open-source frameworks like LiveKit Agents and Pipecat run wherever you deploy them.

Are Vapi, Retell and Bland HIPAA compliant?

Each vendor says it supports HIPAA with a signed Business Associate Agreement. Vapi charges a $2,000 a month add-on for it. Retell and Bland list the BAA among enterprise features. These are vendor statements. Compliance for your use case also depends on how you build and run the agent, so involve your own counsel.

When should I build a voice agent from scratch instead?

When voice is your product, when audio cannot leave your own cloud, or when volume is high enough that platform fees cost more than the engineers who would replace them. For most business uses, such as booking calls or after-hours answering, a platform gets you to a working agent faster and the real work is the call flow and the handoffs.

Related solutions

More comparisons

Guides

Vendor and product names on these pages are trademarks of their owners. Scenox Studios is not affiliated with, endorsed by or partnered with any company named here. Facts come from each vendor's public pages and change without notice.

Tell us about the calls your team can't keep up with.

A 30-minute call with an engineer. We will say which stack fits your call volume and rules, including when a platform alone is enough.

Book a strategy callor email hello@scenoxstudios.com

What happens next

  1. 1

    Strategy call

    We learn how the workflow runs today and what a good outcome would be.

  2. 2

    Workflow audit

    We map the process and measure where a system would save the most time.

  3. 3

    Proposal

    A written scope, timeline and fixed price for your first system. No obligation to build.