Real BYOK pricing math, feature breakdown, and the decision tree we use internally to recommend one over the other.
We build production voice AI on both. This is our honest take.
TL;DR
Best for: healthcare (HIPAA BAA), teams without dedicated AI engineers, projects under 50k min/month, anyone who needs a visual flow editor to build and iterate without code.
Best for: engineering-led teams who want maximum API control, custom SIP/telephony infrastructure, or very high-volume deployments where fine-grained infra tuning pays off.
The Real Pricing Math
Both platforms are BYOK (bring your own keys). Platform fees are just the starting point. Here's a typical stack cost comparison using GPT-4o mini + ElevenLabs Flash TTS + Twilio.
| Cost component | Vapi | Retell |
|---|---|---|
Platform fee Retell includes STT in this fee | $0.05 / min | $0.07 / min |
Speech-to-text (Deepgram) Retell bundles STT; Vapi charges separately | ~$0.008 / min | Included above |
LLM (GPT-4o mini, BYOK) Same cost on both, depends on turn length | ~$0.02–0.05 / min | ~$0.02–0.05 / min |
TTS (ElevenLabs Flash, BYOK) Identical — you bring your own key | ~$0.10–0.15 / min | ~$0.10–0.15 / min |
Telephony (Twilio inbound) Same if using Twilio on both | ~$0.009 / min | ~$0.009 / min |
Effective platform + STT Vapi is marginally cheaper on infra alone | ~$0.058 / min | $0.07 / min |
Typical all-in total Retell slightly better with STT included | $0.23–0.33 / min | $0.21–0.30 / min |
1,000 min / mo
Similar at low volume
10,000 min / mo
Retell ~$200–300 cheaper
100,000 min / mo
Both offer volume discounts — negotiate
Estimates based on 2026 pricing. LLM cost assumes ~150 tokens/turn at GPT-4o mini rates. TTS assumes ElevenLabs Flash Turbo v2. Actual costs vary by call length and provider choice.
Feature Comparison
| Feature | Vapi | Retell |
|---|---|---|
Visual flow builder Retell has drag-and-drop node editor; Vapi requires code | ✗ | ✓ |
API-first design Vapi gives fine-grained programmatic control; Retell is UI-centric | ✓ | ✗ |
HIPAA BAA available Retell offers signed BAA; Vapi HIPAA status is unconfirmed | ✗ | ✓ |
STT included in platform Retell bundles Deepgram; Vapi charges STT separately | ✗ | ✓ |
BYOK LLM Both support GPT-4o, Claude, Gemini, custom endpoints | ✓ | ✓ |
BYOK TTS Both support ElevenLabs, Cartesia, PlayHT, Azure | ✓ | ✓ |
Native analytics dashboard Retell has richer built-in call analytics and transcript search | ✗ | ✓ |
Real-time interruption handling Both handle barge-in; Vapi gives more config control | ✓ | ✓ |
Custom SIP / BYOT Vapi allows full SIP configuration; Retell is more opinionated | ✓ | ✗ |
Multi-language support Both support 20+ languages via provider BYOK | ✓ | ✓ |
Knowledge base / RAG Both support document upload and semantic retrieval | ✓ | ✓ |
Call recording Both record calls; Retell has better native transcript UI | ✓ | ✓ |
Volume Breakpoints
Entry-level projects, pilots, SMB deployments
Enterprise scale — platform choice depends on infra team
HIPAA & Compliance
Retell offers a signed Business Associate Agreement (BAA) and HIPAA-eligible infrastructure. Vapi does not publicly offer a BAA as of 2026. For any healthcare use case — patient intake, appointment scheduling, clinical triage — this is not a close call.
The Decision Tree
This is the actual decision flow we walk clients through. Answer yes to the first question that applies.
Do you need HIPAA compliance or a BAA?
Yes
Use Retell
BAA available, HIPAA-eligible infrastructure
No
Continue ↓
Is your team primarily engineers working in code?
Yes
Continue ↓
No
Use Retell
Visual flow editor — no code required for most agents
Need deep SIP / custom telephony infrastructure?
Yes
Use Vapi
Best infrastructure control for complex SIP setups
No
Continue ↓
Starting volume under 50k minutes/month?
Yes
Use Retell
Lower effective cost per minute at entry scale
No
Continue ↓
None of the above apply?
Yes
Use Retell
Default recommendation — lower total cost and simpler ops
No
Continue ↓
Final default
When in doubt, start with Retell.
Lower total cost, faster to ship, HIPAA ready, and you can migrate to Vapi later if you outgrow it.
Our Role
Hestur AI is a platform-agnostic voice AI engineering firm based in San Francisco. We don't resell Vapi or Retell and have no preferred-partner agreements with either. We recommend the platform that fits your use case, then build on it.
Typically healthcare, insurance, and SMB deployments where a non-engineering team needs to manage the agent without ongoing dev work. We build the initial agent architecture, configure HIPAA-compliant infrastructure, and hand off to your ops team.
Common scope: patient intake, appointment scheduling, inbound triage, customer support deflection.
Typically enterprise clients with an in-house engineering team who need custom telephony infrastructure, deep API integration into existing systems, or multi-platform orchestration across voice, chat, and workflow automation.
Common scope: outbound sales automation, custom SIP routing, multi-step agentic call flows.
Book a 30-minute call. We'll walk through your use case, volume expectations, and compliance requirements — and give you a clear recommendation with no pitch attached.