HEALTHY EXPECTATIONS
If you want to be able to use this app for something other than just fun (especially with your kids), you need to understand what it is at this stage and what it is not. Many of those who have been harsh seem to have very knee-jerk reviews (on AppSumo), and their critiques often reveal they simply did not explore the app very thoroughly, or didn't understand the product's focus, or don't seem to have read any of the many, many, lengthy, respectful and thorough replies the developers have offered for critiques, feedback, hopes, and so forth. It has weaknesses for sure, but it does indeed have some great strengths.
So what is it NOT? Well, it's not a Synthesia or Pippio alternative, nor is it a competitor for Adobe Character Creator. But Puppetry has repeatedly said they are not trying to have the same use cases, so this shouldn't surprise users. It has no rigging system that you can alter (at this stage), and the video-to-video option has yet to arrive on the desktop version, which is (seemingly) contributing to some of the weaknesses.
If you use their Presenter Generator or understand which type of face and view work best, you get solid results, even with more realistic faces than they recommend. Using your own audio ups the success you can have in achieving your desired outcome, though you can find a few very solid text-to-speech options if needed. (With better ones possibly on the way if they end up switching to Elevenlabs).
Speaking of, the text-to-speech is another area people get frustrated, but I don't get the impression they searched the catalog of options very hard, nor do they (seemingly) understand TTS, even when done by ElevenLabs who are the best of the best, still has limits. Narration can be done well, but dialogue is rarely captured skillfully by anyone's TTS, so you have to be smart in your use cases, especially with a talking avatar.
All is not perfect, though, as I mentioned. It would be very helpful if they would bring the video-to-video, as that will hopefully mitigate how sometimes the animation can push itself too far and become warped.
The Presenter Generator works well enough, but I'm looking forward to them expanding the ages and looks options further, as they've discussed doing based on user feedback.
It's also odd how one cannot, no matter what prompt you try, create the ideal view and size for a proper Presenter using their image generation option. (I've used the same prompts to create heads that work great, but I had to use Dall-E2 and upload manually.) Right now, many angles and sizes, hairstyles, backgrounds, etc are problematic, but there is a sweet spot that works well consistently. Their image generator really should have an option to turn on that would keep your results within that range.
The GPT text option seems to confuse people, but since everyone under the sun wants "all-in-one" apps, it makes sense for them to include this for those looking to generate dialogue for the avatar. Perhaps having some directions, notes, or hints to that effect would mitigate those who seem utterly confused by its inclusion? (Or maybe not... people gonna' be people).
They've talked about possibly having the UI show the best alignment, and I think this is a great idea that would help a lot. Something that allows you to align and shape your image to the 'rigging' and produce better results would expedite successful uses and reduce server load with regeneration, trial-an-error.
I'd also love an audio-only download option.
The developers have been super transparent with server upgrades, feature requests, and their roadmap, so they make it easy to root for them, especially given how kind they are to, in my opinion, some unfair reviews. So, I'm enjoying it for what it is right now, have found real use cases for it, and look forward to where it goes.
