Systems Lab
← Back to directory

Retell AI

#1 AI Voice Agent Platform for Automating PHONE Calls

Visit Retell AI →

Retell AI is a platform to build, test, deploy, and monitor AI voice and chat agents, with telephony, prompts, tools, and analytics built in. It targets phone-based use cases like appointment booking, lead qualification, call transfers, and IVR navigation, marketing ~600ms response latency.

Vendor's own claim
activation/deliveryorchestration/agent runtime

Core capabilities

Vendor's own claim
  • Two agent-building modes: node-based conversation flow and single-prompt agents
  • Voice and chat modalities, plus SMS and API access
  • Real-time function calling, appointment booking, lead qualification, call transfers
  • Call center features: warm transfer, IVR, DTMF
  • Knowledge base with website crawling and document upload, auto-synced
  • Batch calling
  • Post-call analysis with sentiment scoring and AI quality assurance
  • Live call monitoring, custom analytics, and A/B testing for agents
  • MCP server to manage agents from Cursor, Claude Desktop, and other MCP clients

API & agent surface

From official docs
Authentication
Bearer token: "Authorization: Bearer <RETELL_API_KEY>", same scheme for REST API and MCP server
Base URL
https://api.retellai.com
Rate limits
Documented only as a 429 "Account rate limited, please throttle your requests" response; specific request-per-minute thresholds are not published in the docs pages checked.
Notable endpoints
  • Create/list/get/delete phone or web calls
  • Create/update/publish/version agents
  • Import/provision/list phone numbers
  • Create knowledge bases, attach/remove sources
  • List voices, clone voices, search community voices
  • Create/assign/terminate chats
  • Create test cases, run QA, submit scores
  • Create/list alert rules and incidents, test webhooks
Machine-readable spec
OpenAPI spec retrieved

Pricing

From official docs
PlanPriceIncludedNotable
Pay-as-you-go$0.07-$0.31/min for voice agents (component-priced: voice infrastructure $0.055/min, TTS $0.015/min, LLM $0.003-$0.16/min, US telephony via Twilio $0.015/min, add-ons +$0.005-$0.01/min each); AI chat agents $0.001-$0.052/message$10 free credits, 20 concurrent calls
n/a
Full platform access, Pre-built agent templates, Call analytics and transcripts
EnterpriseCustom, volume-based50+ concurrent calls, no cap
n/a
24/7 omnichannel support with dedicated portal, Dedicated implementation support, Custom MSA/DPA terms

Checked 2026-08-03 against their pricing page. Vendors change pricing without notice.

What is Retell AI

Vendor's own claim

Retell AI is a voice/chat AI agent platform aimed at automating phone-based workflows for sectors like healthcare, financial services, insurance, logistics, retail, travel/hospitality, and debt collection.

Ideal customer profile

Vendor's own claim

Vendor markets to companies automating high-volume phone workflows - healthcare, financial services, insurance, logistics, retail, travel/hospitality, and debt collection - spanning both startups building on the pay-as-you-go plan and enterprises needing dedicated infrastructure, SSO, and custom terms.

Who should look elsewhere

Our read

Pricing is usage-based and stacks multiple metered components (voice infrastructure $0.055/min, TTS $0.015/min, LLM $0.003-$0.16/min, telephony $0.015/min, plus optional add-ons), so a team wanting a single flat per-minute number for budgeting has to model several line items rather than one price. Charges continue while the caller is silent because speech-to-text stays active, which penalizes use cases with long holds or pauses. Teams above 20 concurrent calls pay an extra $8/call/month for capacity on the pay-as-you-go tier, and anything beyond that (50+ concurrent, SSO, custom MSA/DPA) requires moving to custom Enterprise pricing rather than a published rate.

Retell AI alternatives

Our read

Platforms for building and running conversational voice agents that place or answer phone calls.

Category: Voice AI agents. Membership is decided by 2 stated tests, not by category tags — see them and the full comparison.

Retell AI FAQ

How is Retell AI priced?

Usage-based per minute, billed to the nearest second, built from separately metered components: voice infrastructure ($0.055/min), text-to-speech ($0.015/min for Retell voices), LLM ($0.003-$0.16/min depending on model), and telephony ($0.015/min for US via Twilio), plus optional add-ons. Chat agents are billed per message ($0.001-$0.052).

From official docs
Is there a free tier or trial?

New accounts get $10 in free credits with no commitment required; there is no separate time-boxed trial documented beyond that.

From official docs
Does the bill keep running if a caller goes silent?

Yes - per the pricing page, charges continue during silence because speech-to-text remains active; charges for the AI agent stop once a call is transferred to a human.

From official docs
What happens above 20 concurrent calls?

The pay-as-you-go plan includes 20 concurrent calls free; additional concurrency costs $8/call/month. Enterprise plans start at 50+ concurrent calls with no cap, at custom pricing.

From official docs

Verify before you buy

  • A GitHub org link confirmed from Retell's own site or docs (a plausible org exists but was not reached via a vendor-controlled link)
  • A specific official video URL/title from Retell's YouTube channel (channel link is in the vendor's own footer, but no specific video was verified)
  • Named MCP tool identifiers (docs describe functional areas, not individual tool names)
  • Numeric API rate-limit thresholds (only the 429 error behavior is documented)
  • Named MCP repo or agent-skill repo published by the vendor

Need help setting it up?

This page tells you whether Retell AI fits. Cheetah builds and runs the system it goes into — enrichment, sequencing, CRM and the joins between them.

Book a call →

The directory stays free. There is nothing gated behind this.

Build my GTM engine

Tell us what you want to improve. Fedor reviews the workflow and replies by email when there is a useful fit.

Fedor reviews every request personally. If there is a useful fit, he will reply by email.