AI SmartTalk

Product details
bruce99

Verified purchaser

Deals bought: 47Member since: Apr 2021
2 stars
2 stars
May 26, 2026

Not worth it.

I bought tier 2 wich gives me 2,5 million tokens every month. While the setup of the bot is fairly straight-forward, the token assignment is a joke. My site has not gone live yet, so just installing and testing the bot with less than 10 questions brought my token usuage up to 71,300 tokens used. That is 3% from one chat.

At the rate tokens are taken, not even the tier 4 with 15 Million tokens will work. When my site launches, we will have 10,000 members, probably within the first week. I suspect the token will be finished even before we reach 10,000 members.

I am asking for a refund.

Don't get me wrong, the product is very usefull and if you have a small site with almost no traffic, it will suit you fine, however, anyone with a website that gets a moderate amount of traffic is going to run out of tokens way before the end of the month.

In conclusion, I have to mention that their support is absolutely fantastic and for smaller sites I would definately recommend them. It is actually because of their support that I raised my rating from 1 taco to two tacos.

Founder Team
Odran_AI-SmartTalk

Odran_AI-SmartTalk

Edited May 30, 2026

Hi Bruce, and thanks again for the kind words about our support.

The key update for anyone reading: BYOK is now live. You plug your own model key on the AI Model page of each agent and run on your own provider billing, so the token quota is no longer a ceiling, whatever your traffic. That’s what makes the deal fully viable for high-traffic sites.

On token consumption itself, which is the central point here: usage depends almost entirely on the model selected. GPT-4.1 mini, for example, burns around twice the tokens of GPT-4o mini per conversation for comparable quality on FAQ and support. It’s a one-click change, and the dashboard shows cost per conversation so you can benchmark before going live. Following your feedback we’ve already updated the default models for a more efficient experience out of the box.

Odran

Helpful?
16
Share
Ratings