We use both direct VEO 3 with google tools to build a little working app for video generation using first and last frame, but also FreePik AI Suite, which has access to all video generators, including VEO 3+ If we look at what we have paid for using just these video tools compared to the time cost alone of doing this without AI, its negligible. It is expensive if you don't know what you are doing, so the KEY to this is to fine tune the start and end frames using cheap image generative AIs, like chatGPT or Nano Banana, etc. Galaxy.ai and our personal favorite freepik.com supplies everything you need, and I'm confident (will verify) you pay less per video through freepik than you do directly using your own Veo API due to tiered pricing, where we pay the higher price for tier 1 on our google API. We have generated enough useable animated videos to justify the costs 10 fold. It would have taken us roughly $7,500 in just TIME ALONE to make what we have made so far with VEO3. Hopefuly, it just gets better and cheaper.

In light of Mark’s lifetime of integrity and transparency (two traits I always prioritize in people I trust) — just an FYI for anyone interested. There are a lot of AI platforms popping up right now (browse a directory like “There’s An AI For That” if you want to see the flood). Galaxy.ai is one of the better “all-in-one” places to test a wide range of image/video/audio models in one spot. But it’s worth knowing this: many of these platforms are essentially wrappers around the same underlying tools/models. This feels a lot like an early-stage bubble (think dot-com era). Over the next decade, only the most refined, reliable, and easiest-to-use platforms will really remain. How to think about Galaxy.ai (and platforms like it) Galaxy.ai looks like a safe, convenient way to: - try lots of models quickly - discover what actually fits your workflow - avoid the hassle of wiring everything together yourself - That convenience can absolutely be worth the price—especially while you’re experimenting. If you want to save money later… Once you’ve found the specific models and outputs that match your style, you can often reduce costs by either: 1. going directly to the model/tool you ended up using most, or 2. building your own custom workflow/tool (power-user route) with things like Cursor, Builder.io, or similar “build-your-own app” tools + the model APIs. That’s literally how many of these platforms are made. If Galaxy.ai is perfect for your needs, stay there. If credits start getting expensive (you are a power user or you can't figure it out), graduate to a custom setup once your workflow is proven. Example: one of our current workflows For our children’s illustrated content: - ChatGPT → lyrical structure + prompting - Suno → music - ChatGPT → image frame ideas - Midjourney → visuals/animation direction That workflow works great for us—but it’s not a “cinematic pipeline,” and it can’t fully live inside Galaxy.ai right now because Midjourney doesn’t offer a public API (as of today).

https://marketplace.reallusion.com/

Midjourney has been the most consistent for us to create videos from still images. It is not as powerful as VEO 3.1 or KLING or others, but so far, is the fastest most affordable and consistent. You just have to work on your prompt (chat gpt is helpful with this) so that it doesn't create crazy movements. Made these two vids a a few hours using midjourney, chatgpt, and the music was suno where the narration was elevelabs https://youtu.be/5klkyU9eeHM?si=aW6bkGEjtsYFLKma https://youtu.be/k3WsPLsmztI?si=d7Kt1RIXp_f-l-CB

I decided to post all my content for free some time back: https://marketplace.reallusion.com/cartoon-animator/author/boxels Take any character, export a pose, put it in AI vid generator as 1st frame, make another pose, put it as 2nd frame, rinse and repeat. Easy animation.