I've been using Synthesia for well over a year, pausing my subscription for 2 months when I don't use it. In terms of price, DeepWord is quite interesting. In terms of results it's not quite there.

The actors are stiff : contrary to Synthesia, the body does not move. It looks almost like the head was grafted on a still body image. Head movements are too repetitive and not convincing.

The actors are also very poorly lit, regardless of their ethnicity, their skin color is off. This results is a zombie-like appearance.

To make things worse the "attire" proposed are not even remotely professional. There's a girl wearing some sort of night gown which is described as "casual".
Generally the choice of garment is dismal and the dresses are not even ironed.

Here's a comparison of some of DeepWord's actors https://share.getcloudapp.com/mXuY04Dz
and Synthesia's https://share.getcloudapp.com/lluEPbO1

Synthesia's actors are by no mean perfect but none of them look like a zombie and some look great. None of DeepWord's actors look good.
I don't want any of my client's brands to be associated with a zombie.

To see how both software compare, here's a video created last year with Synthesia (in French) https://www.musicadmix.com/ (scroll down a bit)
and here's the exact same text created with DeepWord : https://login.deepword.co/video/b47y6agycl34o4eqq
Notice how the voice is similar. The Text to Speech feature is fine, and perhaps even better with DeepWord than Synthesia. However, the actor in DeepWord's version is almost scary and makes the whole video much less attractive.

After checking the founder's credential I saw that they did put a lot of work into this and have a good approach. What's funny though is that they describe Synthesia as "puppet approach" when their own product delivers a much less realistic result. But still, the tech is smart which makes me think I won't ask for a refund and I will see where this is heading. One day maybe it will be worth using. But at this stage, it's not even remotely usable for commercial use. Perhaps by filming your own well-lit model it's possible to achieve a better result but I haven't tried that, having precisely zero filming ability.

The rendering time is rather slow (what was described as 10 minutes turned out to be more than 20) and the fact that the green screen background removal tool did not work when I tried the software did not help.

For what it is now I would give 2 stars, but for what it could become I give 4 stars because 3.5 is not possible on AppSumo.

With much better actors, those 4 stars would be fully deserved.