Retell AI vs Vapi

Retell AI

One of the more honest voice-agent stacks: the pricing page breaks out infrastructure, TTS and LLM cost per minute so you can predict a call's cost instead of guessing. Telephony choice is wide and the concurrency allowance is generous. The catch is that per-minute billing plus add-ons makes a busy contact centre expensive fast, and the cheap end of the range only holds if you pick the cheap models.

Pros

• +Per-minute cost broken down by voice infrastructure, TTS provider and LLM on the public pricing page
• +Telephony via SIP trunking, Twilio and Vonage plus n8n, HubSpot, Make and Go High Level integrations
• +20 concurrent calls included before any concurrency charge
• +Voice, chat and SMS agents share one platform and API
• +HIPAA path exists via a signed BAA covered in the terms of service

Cons

• −Cost per minute more than quadruples between the cheapest and most capable model combinations
• −Knowledge base, denoising, PII removal and branded caller ID are all metered extras
• −No flat-rate plan, so a spike in call volume lands straight on the invoice

Vapi

The default answer when a developer needs a voice agent on a phone line rather than a chat widget: it handles the hard real-time parts — barge-in, latency, turn-taking, telephony — and lets you swap in your own models. Enterprise adoption at Ring and Intuit is real validation. Cost forecasting is the hard part, because the platform fee is only one of four meters running.

Pros

• +Handles real-time voice mechanics: turn-taking, interruptions, latency and telephony
• +Bring-your-own-key support across OpenAI, Anthropic and Google avoids double model markup
• +CRM and helpdesk integrations for Salesforce, HubSpot and Zendesk
• +SOC 2, HIPAA and PCI compliance plus SSO and RBAC on the enterprise tier
• +Proven at scale, with Amazon Ring routing inbound support volume through it

Cons

• −The $0.05/minute platform fee sits on top of speech, model and voice provider costs
• −Concurrency beyond ten lines costs $10 per line per month
• −Enterprise compliance features require an annual Scale contract
• −It is a developer platform, so there is no usable no-code path

Is Retell AI cheaper than Vapi?

On published entry prices, Vapi starts lower — $0.05/min against $0.07/min for Retell AI. Entry price is the lowest paid plan as we last verified it; what each plan includes differs.

Does Retell AI or Vapi have a free tier?

Retell AI has a free tier; Vapi has paid plans only.

Which has a public API, Retell AI or Vapi?

Both ship a public API, so either can be built on; the fact table above shows the platforms each one runs on.