Overs
Pricing

Pay for the studio.
Not for the models.

You connect your own OpenRouter key, and every generation goes from your browser to the provider at their list price. We are never in the middle of that call, so there is nothing for us to add to it. What you pay us for is the room: the brands it remembers, the people who can open them, and how much of it runs on its own.

Free

free
  • 40 frames a month
  • One brand
  • One seat
  • No reviewer: you are the first to see each frame
  • Brand Watch when you ask, at most once a month

Pro

$29 a month
  • Unlimited frames a month
  • 3 brands
  • One seat
  • The reviewer checks every frame before you do
  • Brand Watch weekly, on its own

Team

$99 a month
  • Unlimited frames a month
  • 10 brands
  • 5 seats, one workspace
  • The reviewer checks every frame before you do
  • Brand Watch daily, on its own

An extra brand is $10 a month on Pro or Team, added to the same subscription. A frame counts once when it is generated, so a set of four variants is four; frames you already have never expire and are never deleted, whatever plan you are on.

Choose the renderer. Nothing else.

Switchable per run

Rendering is where cost, speed and character genuinely differ, so it is the one decision left to you. Start on Lite while you are finding a direction, then re-run the frames you like on Pro or ChatGPT Image. A single frame can be re-run on its own, so switching is cheap. These prices are the providers’, paid on your key.

Nano Banana 2 Lite

about $0.03 a frame

Quickest and cheapest. Good for finding a direction before you commit.

google/gemini-3.1-flash-lite-image

ChatGPT Image 2.5 Flare

about $0.05 a frame

The quick half of Image 2.5. Everyday frames, about twice as fast as Image 2.

openai/gpt-image-2.5-flare

ChatGPT Image 2.5 Sunburst

about $0.05 a frame

Slowest and most precise. Built for holding a product exactly while the scene changes.

openai/gpt-image-2.5-sunburst

Nano Banana 2

about $0.07 a frame

The everyday choice. Strong on product fidelity at volume.

google/gemini-3.1-flash-image

Nano Banana Pro

about $0.13 a frame

Holds detail and reference likeness best. Slower, and worth it for finals.

google/gemini-3-pro-image

ChatGPT Image 2

about $0.21 a frame

Different eye. Often better on typography, fabric weave and hard studio light.

openai/gpt-5.4-image-2

Any other image model on your account that accepts a reference image also appears in the picker. These four are the ones worth starting from.

The crew is not a choice.

Everything upstream of the render is fixed at the strongest model for its job, on every plan including the free one. The whole thinking chain is a handful of calls, so downgrading it saves pennies and costs you the shoot. Anything that looks at pictures goes to Gemini. Anything that builds structure or checks it goes to Opus.

Reading your images
google/gemini-3.1-pro-preview
Planning the shots
google/gemini-3.1-pro-preview
Art direction
anthropic/claude-opus-4.5
Quality gate
anthropic/claude-opus-4.5

These are preferences, not hard-coded ids. Overs checks your account’s live catalogue when a run starts and falls back to the closest model that can do the job, so a rename upstream degrades a run rather than breaking it.

What a run actually costs

Almost all of it is the images, and none of it is ours. Ten frames runs from well under a dollar on Lite to a few dollars on Pro, and the thinking adds roughly twenty cents either way. Overs shows an estimate before you start, counts the real billed cost as it goes, and shows the final figure against the run.

Worth knowing: that estimate is one render per frame. A frame you actually keep often takes a few attempts, the way a real shoot does, so budget above the first pass rather than at it. That is also why the free plan is forty frames rather than forty keepers.