Overview
Retell AI provides an end-to-end platform and voice API for designing, testing, deploying, and monitoring production-ready AI voice and chat agents that operate over phone, web, SMS, and chat. It supports prompt- and flow-based agent design, simulation testing, observability (call history, success rates, latency, sentiment), and deep integrations with telephony and CRMs. The platform emphasizes low latency, reliability at scale, multi-language support, and enterprise compliance (SOC 2, HIPAA, GDPR). Typical onboarding requires one engineer; enterprise customers can opt for white-glove services.
Key Features
Agent Builder (Conversation Flow & Prompt Engineer)
Create flow-based or prompt-based agents with dynamic variables, multi-prompt, and prompt engineering tools.
Voice & LLM Engine Selection
Choose voice TTS engines (ElevenLabs, OpenAI TTS, Cartesia, etc.) and LLMs (GPT-4.1/4o/5, Claude, Gemini) to balance cost, latency, and quality.
Telephony & Integrations
Native telephony or SIP/Twilio/Vonage integrations, verified numbers, SMS, and international termination support.
Testing & Simulation
Web-based Playground and automated simulation testing to validate and stress-test agents before production.
Monitoring & Analytics
Call history dashboard, call completion and task completion metrics, latency, and sentiment analysis for observability.
Knowledge Base & Functions
KB retrieval, custom functions, webhooks, and realtime data injection to keep agents current and context-aware.
Who Can Use This Tool?
- Developers:Integrate voice agents via API/SDKs and build custom telephony workflows and webhooks.
- Contact Centers:Automate inbound/outbound calling, IVR, and batch outreach to reduce agent workload.
- Enterprises:Deploy compliant, large-volume voice automation with dedicated support and SLA guarantees.
- Product/Operations:Run experimentation, simulation testing, and monitor agent performance for operational efficiencies.
Pricing Plans
Self-serve usage-based plan with free sign-up, trial credits and pay-as-you-go rates.
- ✓Self-serve, start instantly
- ✓$10 free credits on sign-up
- ✓20 free concurrent calls included initially
- ✓10 free knowledge bases included initially
- ✓Support via Discord and Email
- ✓Usage billing: $0.07+/minute for AI voice, $0.002+/msg for AI chat (base indication)
Custom managed plan for large volumes with discounted rates, dedicated support and higher limits.
- ✓Includes everything from Pay-as-you-go
- ✓Discounted pricing based on volumes
- ✓Fully managed agent setup (white-glove optional)
- ✓Additional included concurrency and custom concurrent calls
- ✓Higher knowledge base limits
- ✓Early access to beta features
- ✓Premium private Slack channel support with dedicated teams
- ✓Contact sales / custom pricing (recommended for > $3k/month)
Usage-based voice agent pricing composed of voice engine, LLM and telephony per-minute costs.
- ✓Conversation voice engine: Elevenlabs/Cartesia $0.07/min, OpenAI voices $0.08/min
- ✓LLM per-minute options (examples): GPT-5 $0.04/min, GPT-4o $0.05/min, GPT-4.1 $0.045/min, Claude 4.5 $0.08/min, Gemini 2.0 Flash $0.006/min
- ✓Telephony (Retell Twilio): $0.015/min (country-specific rates available)
- ✓Prompts exceeding 3500 tokens incur surcharge
- ✓Add-ons available (knowledge base, batch calls, branded calls, denoising, PII removal)
Telephony per-minute destination/origination rates for outbound/termination per country.
- ✓US: $0.015/min (Retell Twilio)
- ✓US toll-free: $0.06/min
- ✓India: $0.15/min, Australia/Germany/Spain/UK: $0.1/min
- ✓Japan: $0.28/min, Philippines: $0.8/min, Indonesia: $0.4/min
- ✓Telnyx examples: US $0.03, India $0.25, Canada $0.03
Optional per-use add-ons to augment call agent functionality and pricing.
- ✓Knowledge Base: +$0.005/minute
- ✓Batch Call: +$0.005/dial
- ✓Branded Call: +$0.10/outbound call
- ✓Advanced Denoising: +$0.005/minute
- ✓PII Removal: +$0.01/minute
- ✓Several add-ons have free allowances (e.g., first 10 KBs or concurrency free)
Chat agent pricing charged per AI message with multiple LLM options and rates.
- ✓GPT-4.1: $0.015/AI msg (recommended)
- ✓GPT-4o: $0.017/AI msg; GPT-5: $0.013/AI msg
- ✓Smaller LLMs: GPT-4.1 mini $0.006/msg, GPT-4.1 nano $0.002/msg, GPT-5 mini $0.004/msg, GPT-5 nano $0.001/msg
- ✓Claude and Gemini tiers available (e.g., Claude 3.7 sonnet $0.021/msg, Gemini 2.0 Flash $0.006/msg)
- ✓SMS add-on available ($0.01/Msg) and concurrency/KB free allowances for some tiers
Optional monthly subscription items for phone numbers, SMS, concurrency and knowledge bases.
- ✓Retell Phone Numbers: $2.00/month per number
- ✓Retell SMS: $20.00/month
- ✓Concurrency (active calls): $8.00 per concurrency/month (first 20 concurrency free)
- ✓Knowledge Base: $8.00 per KB/month (first 10 KBs free)
- ✓Verified Phone Number: $100.00 per phone number/month
Pros & Cons
✓ Pros
- ✓Transparent, granular usage-based pricing and pricing calculator for cost simulation
- ✓End-to-end tooling: build, test (simulations), deploy, monitor (analytics) lifecycle
- ✓Enterprise-grade compliance (SOC 2 Type I/II, HIPAA, GDPR) and security focus
- ✓Multiple LLM and TTS engine options to tune cost vs. quality
- ✓Integrations with telephony providers (Twilio, Vonage, SIP), CRMs, and automation tools
- ✓Starter incentives (free credits, free concurrency, free KBs) for quick pilots
✗ Cons
- ✗Cost can become complex to forecast due to multiple per-minute and per-message components
- ✗Telephony international rates vary widely (some high-cost geographies)
- ✗Advanced enterprise features require contacting sales / custom pricing
- ✗Some advanced configuration and integration require engineering resources
Compare with Alternatives
| Feature | Retell AI | Synthflow | TalkStack |
|---|---|---|---|
| Pricing | N/A | N/A | N/A |
| Rating | 8.1/10 | 8.0/10 | 8.1/10 |
| Agent Authoring | Yes | Yes | Yes |
| Model & Voice | Yes | Yes | Partial |
| Telephony & SIP | Yes | Yes | Yes |
| Testing & Simulation | Yes | Yes | Partial |
| Scaling & Batch Calls | Yes | Partial | Partial |
| Compliance & Security | Yes | Yes | Partial |
| Monitoring & Analytics | Yes | Yes | Yes |
| Multilingual Support | No | No | Yes |
Related Articles (2)
GPT-5.1 brings faster, more natural conversations to Retell agents at no extra cost.
Retell AI launches GPT-5.1 with warmer, faster, more capable conversations at the same price.
