Made on YouTube 2025: Dream it, prompt it, create it - now with sound

Create faster on YouTube: Veo 3 with sound, Edit with AI, and Speech to Song

YouTube just rolled out new creative tools built for speed, play, and output. If you make Shorts, you'll get prompt-to-video with sound, auto-edited first drafts, and instant remixes from spoken lines to music. Here's what's live, what's next, and how to use it for real projects.

Veo 3 Fast in Shorts: prompt to video with sound

Veo 3 Fast, a custom version of Google DeepMind's video model, now generates 480p clips with lower latency-free inside YouTube Shorts. For the first time, these AI-generated clips include sound. Access it by tapping Create in the YouTube app, then the sparkle icon to find the latest gen AI tools.

Availability: rolling out in the United States, United Kingdom, Canada, Australia, and New Zealand, with more regions planned. It's built for quick idea testing, mood pieces, and concept shots you can layer into your edit.

• Prompt ideas: "neon-lit city alley, slow camera push-in, rain on pavement, moody synth atmosphere," "stop-motion paper tiger walking across a desk," "cozy kitchen at sunrise, cat yawning on a wooden stool."
• Workflow tip: generate short beats, then stitch with your live footage to keep pace high and style consistent.

New Veo tricks for Shorts: more ways to create with less effort

• Add motion: Apply movement from one video to a photo or another subject. Think dance loops, sports moves, or stylized gestures transferred to your static assets.
• Stylize your video: One-tap looks like pop art or origami to reshape your aesthetic without manual grading.
• Add objects: Insert characters, props, or effects from a text description. Example: "a rubber duck in a coffee mug" or "a giant octopus near the harbor."

YouTube will start experimenting with these on Shorts in the coming months. Plan your prompts and mood references now so you can move fast once they land.

Edit with AI: get your first draft without the blank timeline

Edit with AI turns raw camera roll footage into a ready-to-tweak first draft. It finds strong moments, sequences them, adds music and transitions, and can generate a playful voiceover that responds to the visuals in English or Hindi. This saves time on assembly so you can focus on pacing, hooks, and your signature style.

Currently in testing on Shorts and the YouTube Create app, with expansion to select markets in the coming weeks. Use it to quickly test multiple cuts, then iterate on the version with the strongest 3-second hook.

• Fast workflow: import → AI draft → punch up the hook → refine text overlays → color pass → publish.
• Keep control: swap the AI voiceover with your own track and adjust beats to match your brand tone.

Speech to Song: turn dialogue into a soundtrack

Speech to Song lets you remix dialogue from eligible videos into catchy music using Lyria 2, Google DeepMind's latest AI music model. Choose a vibe like chill, danceable, or fun, and it will transform spoken lines into a melody and rhythm for your next Short. The final result attributes the original creator.

Use it to spin quotable lines into hooks, flip behind-the-scenes chatter into earworms, or give your GRWM a theme that sticks. It's a quick way to create audio identity directly from your content.

AI transparency: watermarks and labels

AI-generated outputs include SynthID watermarks and content labels so viewers know how content was created. Learn more about the watermarking approach here: SynthID by Google DeepMind.

How to get started today

• Open YouTube → Create → tap the sparkle icon → try Veo 3 Fast for a sound-on concept clip.
• Draft a Short with Edit with AI in YouTube Create, then refine your hook, captions, and pacing.
• Test Speech to Song on eligible dialogue to build a repeatable audio motif for your series.
• Batch ideas: generate 5-10 micro-concepts, publish the top 3, double down on the best performer.

Prompt formulas that work

• Look + Action + Mood + Camera: "paper-craft whale breaching, serene, soft morning light, slow dolly."
• Style + Texture + Motion: "pop art halftone, bold primaries, quick zooms and jump cuts."
• Object add-on: "add floating neon arrow pointing to the latte art."

Creative guardrails

• Keep it short: generate 3-6 second beats and stack them for rhythm.
• Match your brand: reuse palettes, typography, and transitions across edits.
• Respect rights: use Speech to Song only on eligible videos; attribution is included on outputs.
• Iterate fast: publish, review retention, tweak the first 3 seconds, and retest.

Want structured training to sharpen your prompt craft and video workflow? Explore curated tools and training for generative video here: Generative Video Tools - Complete AI Training.