Retell AI Reviews & Product Details

The modularity and ease of use. They have inbuilt phone number thing/ Review collected by and hosted on G2.com.

Well there are some bugs, when I use it for building agent, the workflow was made and when i clicked a function or smth. all the workflow was gone! also there is no straight line when it comes to the pipelining. it is always a bit moved to off side, as a person who has OCD i didnt like that and it does bug me. Sometimes when i am dragging it as a pan, it directly starts to select rather than moving to change the POV. they should fix it. Review collected by and hosted on G2.com.

I like to you can add multiple integration programs and use different AI voices Review collected by and hosted on G2.com.

It's a bit hard to find support when you have a question. Review collected by and hosted on G2.com.

What I like most about Retell AI is the low-latency conversational performance and how easy it is to build production-ready voice agents. The call quality feels natural, the APIs are developer-friendly, and integrations with LLMs, SIP, and telephony providers are straightforward. It significantly reduces the complexity of building AI phone systems while still allowing deep customization for workflows, prompts, and call handling logic. Review collected by and hosted on G2.com.

One downside of Retell AI is that advanced configurations can take time to optimize, especially for complex conversational flows and edge cases. Pricing can also become expensive at scale depending on call volume and model usage. Additionally, some areas of the documentation could be more detailed for troubleshooting and enterprise-level deployment scenarios. Review collected by and hosted on G2.com.

More settings about voice. I think the voice in an agent would be better. The UI/UX is so impressive. The Integrations with Netelip i think is correct, but not's easy. The Performance is incredible, my client is so happy with taht service. About the pricing i think is the best of Retell AI, i love the model "Pay as go". With module of support and onboarding I think is fast. And about AI i think Retell have the best models in the market. Review collected by and hosted on G2.com.

I like everything about it. I think the application works correctly. Review collected by and hosted on G2.com.

The voice quality is the standout — it sounds natural and doesn't immediately tip off callers that they're talking to a bot. The dashboard is clean, setup is straightforward, and building out call flows didn't require a ton of technical overhead. Good foundation for anyone running AI voice workflows Review collected by and hosted on G2.com.

Pricing can add up quickly depending on call volume, so it's worth keeping an eye on usage. Would love more built-in analytics to track call performance without needing to piece things together manually. Review collected by and hosted on G2.com.

The ultra-low latency and smoothness of the conversations are impressive — patients don't realize they are talking to an AI agent. The integration with ElevenLabs for voices and compatibility with Telnyx for Swiss local numbers make deployment in a medical office truly viable. The dashboard is clean and real-time call monitoring is a real plus for the agency. Review collected by and hosted on G2.com.

The documentation lacks depth on certain edge cases (handling long silences, multiple interruptions). Pricing can become high as call minutes scale. There are some sporadic bugs in detecting end of turn in French — the English language clearly remains a priority in optimizations. Review collected by and hosted on G2.com.

Best-in-class low latency (~600ms)

This is the standout feature. Conversations feel shockingly natural — quick responses, excellent interruption handling, and smooth turn-taking. It’s currently one of the most human-like experiences you can get in production voice agents.

Realistic voices + strong interruption handling

Voices (especially when paired with ElevenLabs or similar) sound very natural, and the agent handles people cutting in or talking over it without breaking.

Developer-friendly with great flexibility

You can mix different LLMs, voice providers, and telephony (Twilio, Telnyx, SIP trunking, etc.). The visual flow builder + code nodes gives you both no-code speed and deep customization.

Real-time function calling & integrations

It shines when booking appointments, checking calendars, updating CRMs, or pulling live data during the call.

Transparent pay-as-you-go pricing + generous free tier

$0.07+/min + $10 free credits and 20 free concurrent calls makes it easy to test and scale without big upfront commitments.

Strong for real business use cases (especially dental/medical, support, outbound) Review collected by and hosted on G2.com.

Steep learning curve for non-technical users

While it has a visual builder, building truly reliable agents still often needs prompt engineering and some technical knowledge.

Unpredictable total cost at scale

You pay separately for LLM, STT, TTS, telephony, etc. The final per-minute cost can climb higher than expected if you're not careful.

Support can be slow

Many users report that customer support is limited (mostly Discord/community), especially for urgent production issues.

Still requires work for complex flows

It’s not fully “set it and forget it.” You’ll spend time testing and refining to handle edge cases well.

Limited native phone numbers

For international or certain regions, you often need to bring your own carrier (like Twilio). Review collected by and hosted on G2.com.

I love how there are not many companies that allow users to do what Retell does. By this, I mean I have used the tool so much that I can literally create a functional AI Voice Agent/ Receptionist for anyone or any business in as little as 30 minutes. Review collected by and hosted on G2.com.

I understand you're still refining things on your end, but the phone number verification process was a bit confusing and long. As a small business owner, I understand that it takes time for processes to be refined, so this was nothing big enough to deter me from using RetellAI.

Still, I also want to state that I have already created phone numbers (and will soon create separate accounts) for about 5 of our clients, with more on the way soon, and it would be nice to know that the phone number verification process is quicker now than it was earlier this year. Review collected by and hosted on G2.com.

How realistic the conversation's are, the voices are truely great. I useed SignalWire, and Trilio, and Retell is the best as far as I am concerned! The Support team is 24/7 and are equally amazing. Highly recommend and I don't recommend many products. I don't have time, and most aren't worth a review. I never heard of Retell, but I am glad I stumbled upon them and now I'll help promote them to anyone looking for a truely human voice agent. Review collected by and hosted on G2.com.

Nothing, great product! Once you get your agent dialed in, it works amazing! could use more Voices? But I found the ones that work for my business, with the selection they have. Too many and you spend days looking for a perfect voice. I think I like the less is more, I know it's faster and easier the way you have it now. Review collected by and hosted on G2.com.

The conversations don’t feel robotic like most AI calling tools. The voice responses sound natural, the response time is fast, and it can handle real conversations smoothly without awkward pauses.

I also liked how easy it was to build and customize call flows based on different business needs. Most of the important features are already available within the platform itself, so there’s very little extra configuration or setup needed compared to other tools. The integrations and automation features save a lot of manual work, especially for customer support and follow-up calls. Overall, it feels like a practical, complete, and reliable platform rather than just another AI demo tool. Review collected by and hosted on G2.com.

while the platform is powerful, some advanced customizations and workflow debugging can take time to fully understand in the beginning. There’s a learning curve when trying to fine-tune conversations for more complex use cases. Review collected by and hosted on G2.com.