|
•••
|
|
Google recently changed the math on AI video. Veo 3.1, the model that competes directly with Sora, quietly moved from the paid tiers down to the free tier of Google's AI ecosystem. No subscription required. No card. No waitlist. Just a Google login and access to the same engine paying users were running the day before.
It barely registered as news. Sora's shutdown a few weeks later took up most of the AI video coverage, and Google did not put much marketing behind the announcement. Most creators still have not opened the tool. The math works out to about two AI videos per day, every day, which lands at sixty-plus videos per month if you log in regularly. Seven on your first day thanks to a one-time bonus credit. Two per day after that, indefinitely.
That is not a trial period. That is the standing offer. Same model the paid users get, with three caveats around resolution, watermarking, and licensing that we will get to. Here is everything you need to start using it today.
|
|
One
What Google actually did.
|
|
Recently, Veo 3.1 moved out of the Pro and Ultra tiers and into open access for any Google account holder. Same model. Same quality ceiling. Same eight-second clips. The differences from paid are a visible Made-with-Veo watermark on outputs, capped resolution at 720p, and personal-use licensing instead of commercial rights. Everything else is the model paying users were running last month.
The model itself is the one most analysts consider the strongest video generator outside the ByteDance and Alibaba labs. It competes directly with Sora, which OpenAI shut down shortly after Google made theirs available without payment. The contrast is worth sitting with. The model that defined the AI video category for most creators is gone. The model that quietly went open-access is still here, still updating, and most creators have not opened it yet.
|
|
Two
Where the videos actually come from.
|
|
Veo 3.1 lives in two places, and they serve different jobs.
Google Flow is the filmmaking interface. Built for cinematic generation. Aspect ratio control, character references, multiple variations per prompt, and the kind of setup that lets you direct a scene rather than guess at a clip. This is the tool to use for narrative, atmospheric, or visually ambitious work. Access it through the Google Labs portal under Flow.
Google Vids is the video editor. Veo 3.1 is built directly into the timeline alongside the rest of the editing tools. Type a prompt, get an eight-second clip in your project, drop it next to your other footage. This is the tool to use when the AI clip is one element of a larger video, not the centerpiece. Access it inside Google Drive next to Docs and Sheets.
Most creators will use both. Flow for the cinematic showpieces. Vids for b-roll, transitions, and finishing pieces inside an existing project. Same Veo 3.1 model under both interfaces, different containers around it.
|
|
Three
The credit math, in plain language.
|
|
Flow operates on daily credits. Free accounts get fifty credits per day, with a one-time bonus of one hundred credits on first use. Each Veo 3.1 video costs twenty credits to generate. Credits reset at midnight Pacific time.
The math: on day one you have one hundred fifty credits, which generates seven videos. On day two your credits reset to fifty and the bonus is gone, so you can generate two videos. Same on day three, day four, day five. Two videos per day, every day, indefinitely.
Across a typical month: seven videos on the first day, plus two videos for each of the remaining days. Sixty-five-ish per month is the realistic number if you log in daily. Even at a half-rate cadence — say, you only remember to log in three times a week — you still land north of thirty videos a month at no cost.
In some countries the daily allotment is one hundred eighty credits instead of fifty, which raises daily output to nine videos and pushes the monthly number well past two hundred. Worth checking what tier your country is on. Either way, the floor is generous enough that almost no creator runs out of room before they run out of ideas.
|
|
Four
What you get and what you do not.
|
|
What this gives you. Eight-second clips at 720p resolution. Veo 3.1's full quality ceiling at that resolution, which is genuinely cinematic. Native audio in Flow including ambient sound and dialogue. Image-to-video animation, where you upload a still and describe what should happen in it. Cinematic prompt understanding — the model knows shot terminology, lighting language, and camera movement vocabulary.
What it does not give you. Anything above 720p (paid tiers go to 1080p). Watermark removal — only Ultra at $249.99 per month removes the Made-with-Veo mark in the corner. Commercial licensing for large-scale productions (personal use only at no cost). Clips longer than eight seconds (this is a model limit, not a tier limit, and it applies to paid users too).
The watermark is the constraint most creators feel first. It sits in a corner of every clip and is fully visible. For social posts where production polish is part of the brand, this matters and you may need a paid tier. For drafts, storyboards, mood pieces, personal projects, and most internal use, the watermark is a non-issue and open access covers everything you need.
|
|
Five
How to get the most from each generation.
|
|
Veo 3.1 responds dramatically better to cinematic prompts than to keyword lists. The single best change you can make is structuring your prompt as a director's note rather than an adjective string.
Director's-note format. Shot type first, then subject and action, then lighting and mood, then camera movement and lens feel. A prompt like “a woman on a windowsill, golden hour, cinematic” produces an okay clip. A prompt like the one below produces something the model can actually render with intention:
Slow dolly shot. A woman in her thirties with short dark hair sits on a sunlit windowsill, looking outside at a garden. Warm afternoon golden hour light filtered through linen curtains. Shallow depth of field, thirty-five millimeter lens feel. Ambient sound of distant birds and light wind through leaves.
Image-to-video gets you significantly more visual control than text alone. Generate or upload a still image first, then describe what you want to happen in it. The model preserves composition, color palette, and character details from the reference, which makes the output far more predictable.
Two videos per day is enough to be picky. Spend the credits on prompts you have actually thought through. Vague prompts produce vague clips and the credits are gone either way. Write the full prompt out first, paste it once, and use the second daily generation to refine rather than guess.
|
|
•••
I am putting together a Veo 3.1 prompt pack. Fifteen tested prompts across cinematic, kinetic, product, and explainer categories. Each one in director's-note format, each one ready to paste into Flow.
Want it when it ships? Reply with send me the Veo prompt pack and I will get it to you.
|
|
A QUESTION FOR YOU
Have you opened Google Flow yet, or is this issue the first time you are hearing about it?
Reply and tell me. The first-time openers get a starter prompt I will write for whatever scene they want to test.
If this issue resonated, forward it to a creator who has not opened Flow yet.
|
|
Until next time,
Luxe Prompting
|
|
Luxe Prompting
AI Image Generation for Creators
|
|
