Google Veo 3.1: New Ingredients Make Consistent AI Characters! + Huge Price Drop

I’ve been waiting for a solution to the character consistency problem in AI video, and Google just shipped one. The new Veo 3.1 update introduces what they call “ingredients” — basically reference images you can feed the model to keep faces, outfits, and backgrounds consistent across multiple clips. If you’ve ever tried to build a narrative or even just a cohesive social media series with AI video, you know how frustrating it is when your main character looks completely different in every shot.

What makes this update worth covering is that it actually works, and Google dropped the price by nearly a third at the same time. That combination makes high-quality AI video generation feel accessible in a way it hasn’t before. I wanted to test it immediately, so I recorded a full walkthrough showing exactly how to use these ingredients in practice.

How Google Veo 3.1 Ingredients Solve Character Consistency

The video walks through the Veo interface inside Google Docs, which is where you access the tool right now. I show side-by-side comparisons of videos generated with standard text prompts versus videos that use ingredient-based references. The difference is striking — when you upload a reference image of a face or a specific background, Veo maintains that visual identity across generations. You can stack multiple ingredients too, so you can lock in a character’s appearance and the environment they’re in at the same time.

I also cover the new native vertical output option, which is a small but meaningful addition if you’re making content for YouTube Shorts or Instagram Reels. No more cropping in post. The pricing breakdown is included as well, because that context matters when you’re deciding whether to experiment with a new tool or commit to using it in production.

Watch this and more on the Agent Baltic YouTube channel. For free and paid resources, visit https://payhip.com/agentbaltic.