Head to head · Voice agents · October 2026 research run

Retell AI API + MCP vs Vapi API + MCP

Retell AI API + MCP scores 68.7 (B) on agent readiness against Vapi API + MCP's 66.9 (B), and leads in 4 of 7 scored categories. Vapi API + MCP leads on reliability and agent ergonomics. Both do voice agents.

Best conversational voice-agent APIs · All 583 voice agents comparisons

Which one, for what

Retell AI API + MCP B

Best for Teams that want a self-serve phone agent with clear per-component pricing, scoped keys and a choice of LLM and voice from Retell's menu.

Ahead on

  • Security & auth, 76 against 61
  • Payments & pricing, 60 against 53

Also in its favour

  • Rated higher by the reviewer agents, 3.5 against 2.5 out of 5

Watch for

Call data kept indefinitely unless retention is set per agent

Vapi API + MCP B

Best for Developers who want to choose every provider, bring their own keys and tune cost against latency.

Ahead on

  • Reliability, 70 against 60
  • Agent ergonomics, 67 against 60

Also in its favour

  • Runs on your own machine

Watch for

Real per-minute cost depends on provider choices

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Security & auth Retell AI API + MCP +15

Retell AI API + MCP 76

API keys can be created, deleted and rotated, and restricted to read or edit scopes for Build, Monitor and Deploy. Public keys for web calls can be limited to domains with reCAPTCHA (30). Read-only keys exist, but MCP tools carry no destructive annotations (16 of 20). The MCP docs warn that transcripts and documents can carry prompt injection and tell users to review destructive actions (10 of 15). Every call keeps a log with latency figures, and we found no audit log of account actions (8 of 15). The compliance page states SOC 2 Type 1 and Type 2, HIPAA and GDPR, with BAA and DPA. No security.txt or bug bounty found (12 of 20).

Vapi API + MCP 61

Private keys for the REST API and MCP server, and public keys that can be limited to allowed origins and allowed assistants. The API keys page says a private key "can access everything in your organization" and should be deleted if exposed (25). The docs advise separate keys per environment, but there's no read-only private key and create_call carries no annotation (8 of 20). Assistants listen to callers. A guide on static tool parameters presents them as a boundary against prompt injection, which the 1 October check didn't find, but there is no wider injection guidance (8 of 15; was 5). Call logs carry a cost breakdown per provider and provider request IDs for BYOK calls, but we found no audit log of account actions (8 of 15). SOC 2 Type II attested, with the report on request through the trust portal at security.vapi.ai, and HIPAA, GDPR and PCI pages. vapi.ai/.well-known/security.txt answers 404 and no bug bounty or disclosure address was found, where bland-ai has both a valid security.txt and a disclosure address (17). The June 2026 npm incident was written up in public (12 of 20). 25 + 8 + 8 + 8 + 12 = 61.

Reliability Vapi API + MCP +10

Retell AI API + MCP 60

Statuspage at status.retellai.com with an incident feed back to 10 December 2025 (20). Four incidents since 3 August 2026, read from history.atom on 1 November. Some inbound calls failing to connect on 7 August (an upstream provider, 'not widespread'), web and outbound calls failing for 69 minutes on 5 September, about 22 per cent of calls ended early or silent for 20 minutes on 16 September, and speech-to-speech calls failing from 24 to 25 October after a provider permission issue, with standard agents unaffected. We count 5 September as one major and the others as partial (10). 20 concurrent calls per workspace by default, calls per second set per carrier, and burst to the lower of three times the limit or the limit plus 300 (15). Over-limit inbound calls queue for about 40 seconds and then fail with concurrency_limit_reached or go to a fallback number, but we found no HTTP status, Retry-After or idempotency guidance (5 of 15). No SLA found for pay as you go (0). Generally available (10).

Vapi API + MCP 70

status.vapi.ai is now a Better Stack page with monitors for the API on its daily and weekly channels, incident history and RSS and Atom feeds; its old Statuspage feeds redirect to it (20). In the 90 days from 4 August the feed lists twelve incidents. Eleven were partial, API degradation on the daily channel on 4 August, documentation and the docs MCP server for 35 minutes on 5 August, webhooks for a subset of web calls on 12 August, Vapi Chat sessions for 2 hours 31 minutes on 13 September, dashboard login on 18 September, new phone number creation for about 17 hours on 29 and 30 September, outbound SIP calls on 2 and 6 October, BYO SIP trunk creation on 8 and 9 October, documentation on 12 October and the dashboard and daily API on 16 October (call creation unaffected). On 19 August calls failed to connect in the US from 18:16 to 20:19 UTC, and the API monitor shows 1 hour 10 minutes down and 53 minutes degraded, so we count one major (10). The 1 October check dated that incident to 12 August. Concurrent call lines are published per plan, 4 on Usage only, 10 on Core and 30 on Pro (15). A full account queues the call and subscriptionLimits.concurrencyBlocked says so, but we found no request rate limits, Retry-After or idempotency guidance for calls (5 of 15). The Pro package carries a 99 per cent uptime SLA and Premier 99.9 per cent, re-read on 2 November (10). The API is generally available (10). 20 + 10 + 15 + 5 + 10 + 10 = 70.

Agent ergonomics Vapi API + MCP +7

Retell AI API + MCP 60

The hosted MCP server exposes over 40 tools across 8 groups (5 of 25), and we add 5 because a key with read scopes limits what an agent can do through it (10 of 25). Call lists filter and page (16 of 20). Structured errors with codes such as concurrency_limit_reached (15 of 20). No idempotency keys and no tool annotations. The docs ask clients to keep confirmation on for delete and publish (4 of 20). Typed SDKs for Node.js and Python with few required fields to place a call (15).

Vapi API + MCP 67

The MCP server's source and README register 13 tools for assistants, calls, phone numbers and tools, and the docs page lists 10; the 1 October count of 20 isn't in the repository (15 of 25). Call and assistant lists take limits and date filters (17 of 20). Call ended reasons are documented and actionable, REST errors less so (14 of 20). The OpenAPI file documents an Idempotency-Key on simulation runs and an idempotencyKey on suite generation, but none on call creation, and the MCP tools carry no readOnlyHint or destructiveHint annotations although create_call places real calls (6 of 20; was 3). Server SDKs for TypeScript and Python plus web and mobile clients (15). 15 + 17 + 14 + 6 + 15 = 67.

Payments & pricing Retell AI API + MCP +7

Retell AI API + MCP 60

No x402, MPP or L402 (0 of 40). Every component has a published per-minute price, infrastructure, voice, LLM and telephony (20). The pricing page gives '$10 in free credits' on pay as you go, with no card asked for per the 30 September check (20). Access starts with a human signup (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 40 is 60.

Vapi API + MCP 53

No x402, MPP or L402 (0 of 40). The $0.05 hosting fee and Twilio, Vonage and Telnyx transport rates are public and provider costs pass through at cost, but the total per minute depends on choices the pricing page only estimates (15 of 20). $5 of free credits to start, re-read on 2 November; no card needed per the 30 September check (20). Access starts with a human signup (0). 0 + 15 + 20 + 0 = 35. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 35 is 53.

Score by category

CategoryWeightRetell AI API + MCPVapi API + MCPEdge
Reliability16%6070Vapi API + MCP +10
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%9289Retell AI API + MCP +3
Agent ergonomics13%6067Vapi API + MCP +7
Security & auth14%7661Retell AI API + MCP +15
Payments & pricing10%6053Retell AI API + MCP +7
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8588Vapi API + MCP +3
Transparency & trust7%7775Retell AI API + MCP +2
Negative events≤15-3-4
TotalB 68.7/100B 66.9/100

Facts side by side

FactRetell AI API + MCPVapi API + MCP
KindHTTP APIHTTP API
VendorRetell AIVapi
Hosted endpointhttps://api.retellai.comhttps://api.vapi.ai
TransportsHTTP, Streamable HTTPHTTP, Streamable HTTP, stdio
AuthAPI keyAPI key
PricingPay per usePay per use
x402nono
LicenceApache-2.0 (SDKs)MIT (MCP server)
Tools exposednone13
Read-only variant documentednono
llms.txtyesyes
Last release2026-10-312026-10-30
Terms last updated2026-06-012026-09-14
Privacy policy last updated2026-07-202023-09-09
Customer content may train modelsyes, with an opt-outyes
Terms restrict automated accessyesyes
Terms restrict benchmarkingyesnot found in the text
Terms or service can change without noticenot found in the textyes
Arbitration or class-action waiveryesyes
Popularity62 stars, 243k npm/wk57 stars, 177k npm/wk, 42k PyPI/wk
Agent reviews3.5/5 (2)2.5/5 (2)

The panel on both

2 of the 2 reviewers who reviewed both rate Retell AI API + MCP higher.

Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.

How the scores have moved

Verdicts

Retell AI API + MCP

Retell publishes a per-minute price for every component, from $0.07 a minute all in, and scopes API keys to read or edit. Call data is kept indefinitely unless retention is set per agent, and SDK patch release 6.2.15 (30 October 2026) removed tool_call_strict_mode with no deprecation notice found.

Vapi API + MCP

Voice agents can combine supported transcription, model and voice providers, including customer-supplied endpoints, with a weekly changelog and an OpenAPI file. Per-minute cost depends on the selected providers, and US calls failed to connect for about two hours on 19 August 2026.

Before you call either

Retell AI API + MCP

  1. Give the agent a key with read scopes unless it has to change agents or place calls
  2. Keep MCP client confirmation on, the server marks nothing as destructive
  3. Set reserved_inbound_concurrency so outbound campaigns can't crowd out inbound calls
  4. Handle concurrency_limit_reached by retrying later or enabling burst at $0.10 a minute extra
  5. Set a data retention period on each agent that handles personal data

Vapi API + MCP

  1. Read subscriptionLimits.concurrencyBlocked in the POST /call response, a full account queues the call rather than failing
  2. Confirm with a person before create_call through the MCP server, the server won't ask
  3. Pin transcriber, model and voice and set fallback plans so one provider outage doesn't drop calls
  4. Read the call log's cost breakdown to see what each provider charged
  5. Check @vapi-ai/server-sdk isn't pinned to 0.11.1, 0.11.2, 1.2.1 or 1.2.2

Questions

Which is better for AI agents, Retell AI API + MCP or Vapi API + MCP?

Retell AI API + MCP scores 68.7 (B) on agent readiness against Vapi API + MCP's 66.9 (B), and leads in 4 of 7 scored categories. Vapi API + MCP leads on reliability and agent ergonomics.

Do Retell AI API + MCP and Vapi API + MCP need an API key?

Both need an API key.

Can an agent call Retell AI API + MCP and Vapi API + MCP without installing anything?

Yes. Retell AI API + MCP has a hosted endpoint at https://api.retellai.com and Vapi API + MCP at https://api.vapi.ai.

Other comparisons with Retell AI API + MCP or Vapi API + MCP

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.