Credits are spent by AI and rendering work, not by everyday actions like browsing your catalog, editing a template by hand, or viewing a channel. Here's what actually draws from your budget.
What uses my credits
Every AI and rendering action in Renable spends credits — here are the usage types the Billing page breaks your spending into, what each one costs, and how to check the estimate before you run something at scale.
Updated September 21, 2026
The usage types#
Every credit charge is recorded under a usage type — the kind of work it paid for. These are the names you'll see on the Billing page's Credit usage chart, and they are the same names throughout the product.
Most usage types also record a subject: the thing the work was done for. That is a prompt, a channel, a catalog, or a template — something you set up. It is what lets you see which prompt, which catalog or which channel your credits are going to, not just which kind of work. Where a model is what varies, the model is part of the usage type instead, so a subject always names one of your own things.
| Usage type | What it is | Subject |
|---|---|---|
| Templated images | Rendering a Studio template into an output image, for a channel or for a Studio preview. This runs automatically whenever a product or template changes, and again for every size the template supports. | The channel (or the template, for a Studio preview) |
| Templated videos | The same, for video output. | The channel (or the template) |
| Prompt-written text | A prompt that writes text for your products — descriptions, headlines, reason-to-buys, product titles, voice scripts. | The prompt |
| Prompt-generated images | A prompt that generates an image per product. | The prompt |
| Prompt-generated videos | A prompt that generates video per product. | The prompt |
| Background removal (v1), (v2), (v3) | Cutting the product out of its background, as part of enrichment. Each version is its own usage type, so you can see which one your credits went on — they are the same v1, v2 and v3 you pick as a Background remover in a catalog's image processing pipeline, and they cost very different amounts. | The catalog |
| Image categorization | Sorting product images into categories like product, lifestyle and information. | The catalog |
| Background removal quality check | Verifying that a background removal really produced a transparent image. | The catalog |
| Review quotes | Pulling quotable fragments out of your connected review feeds so they can be used in ads. | The catalog the reviews came from |
| Aimée conversations | Chatting with Aimée anywhere in the product. | — |
| Aimée conversational image generation | An image Aimée creates for you on request in Studio — one image, from a conversation, rather than one per product. | The generation model |
| Aimée conversational video generation | The same, for video. | The generation model |
Together that is fourteen usage types. Usage recorded before these types existed appears under a broader set of names — AI assistant, Template rendering, Image processing, Prompts, Review processing, Copy writing and Image captions. Nothing spends credits under those any more, and the chart groups them separately so they are never compared with the types that replaced them.
Regular editing — moving elements around, adjusting placeholders, changing settings — doesn't consume credits. Only the AI and rendering steps do.
Credit costs by operation#
Approximate costs, per 1,000 products, for a single unit of each operation (one image, one prompt run, and so on) against a cache miss. These are operations rather than usage types — one usage type can cover several of them, and what a given operation costs depends on the model behind it:
| Operation | ~Credits / 1,000 |
|---|---|
| Ad-template image render | ~0.65 |
| Ad-template video render (10s) | ~20 |
| AI text prompt (each) | ~5 (a full title + headline + copy + reason-to-buy suite runs ~20; fact-checking roughly doubles it) |
| AI lifestyle image generation | ~104 |
| AI: animate a lifestyle image into video | ~325 |
| AI freestyle text-to-video generation (8s) | ~10,400 |
| Background removal (v1 / v2 / v3) | ~5 / ~1,820 / ~15 |
| Image classification | ~1 |
| Review-quote extraction | up to ~20 |
A few things to keep in mind when reading these numbers:
- Video generation is by far the most expensive operation — especially freestyle text-to-video. Animating an existing lifestyle image into video is much cheaper than generating video from scratch, but still far more expensive than a still image.
- The background remover versions are not close in price. v2 is the highest-quality one and costs around 120 times what v3 does, and around 360 times v1. Because each version is its own usage type on the Billing page, you can see at a glance which one a catalog is running and what it is costing you.
- These costs are charged only on a cache miss — once Renable has generated a result for a given input, reusing it doesn't spend credits again.
- Prompt runs and Aimée's image and video generation are charged when the work starts, not when it finishes — the three Prompt-… usage types and the two Aimée conversational … ones. A generation that fails part way through still costs credits, and a generation that keeps failing is charged for every attempt, not just the one that eventually works. Renable can't ask a model to undo work it has already done, and charging up front is what stops a single broken product from retrying forever and quietly spending a month's budget.
- Everything else is charged once the work succeeds, so a failure there costs nothing: templated image and video renders, all three background removers, image categorization, the background removal quality check, review quotes, and Aimée conversations.
- Your real total scales with how many units each product needs — how many images per product go through background removal, how many text prompts you run, and so on. A prompt that touches five fields per product costs roughly five times what a single-field prompt costs.
Check the cost before you commit#
You don't have to guess. In the Prompt workbench, running a prompt against a single product with Generate gives you a live preview of the result — and once it completes, Renable shows an estimated cost to apply that prompt to every product in its scope, along with your remaining credit balance. Use this to sanity-check cost before rolling a prompt out broadly, especially for image- and video-generation prompts.