Should Video Generation Be Metered by Seconds or Projects?

Why video duration is the clearest underlying usage unit

Video generation has a variable cost structure. A two-second product animation and a 20-second ad may use the same creative workflow, yet the longer output consumes far more model time. Duration gives product teams a simple, visible unit for that difference.

For a SaaS product, per-second metering makes customer usage easier to explain: customers pay for the seconds of video their requests generate. That connection is especially useful when customers can choose clip length, create multiple variations, or use several models with different generation costs.

Consider a commerce platform that lets merchants turn catalog images into short product clips. A merchant producing a six-second clip for a product listing has a small, understandable usage event. A larger seller creating 30 clips for a seasonal campaign creates a larger usage event. Seconds provide a direct way to measure both.

Protoface API is billed per generated second, giving teams a direct usage basis when they add hosted AI video generation to an ad tool, UGC app, or creative SaaS product. That basis can be passed through to customers as usage credits, included seconds, or overage billing.

Where project bundles hide expensive generation behavior

Project pricing works well when a project has firm boundaries. A product may define one project as a single campaign with one selected model, a fixed clip length, and a limited number of outputs. In that case, a bundle gives customers a predictable purchase decision.

Problems appear when “project” can mean almost anything. One customer may generate a single five-second clip. Another may create dozens of 15-second variations, switch models, add revised prompts, and export multiple formats. Both may call the work one project, while their underlying generation costs differ sharply.

Bundles can still be useful as a packaging layer. They work best when the product defines included usage clearly:

• A maximum generated duration per project

• A set number of generation attempts

• A defined model tier

• A clear charge for additional seconds

This approach keeps the sales page simple while preserving a workable cost model. The key is making the included generation allowance visible before a customer starts creating.

How retries and drafts affect customer billing expectations

Video generation is iterative. Customers test prompts, swap reference images, adjust camera direction, and request another take when a result misses the mark. A billing model needs to account for that behavior without making experimentation feel risky.

Per-second metering handles drafts cleanly when each request shows its expected duration and resulting usage. Customers can see that three eight-second drafts consume 24 generated seconds, even if they publish only one of them. That matches how the underlying service is used.

Your product should also decide which events count as billable. A generation request that successfully produces a clip is straightforward. Provider errors, technical failures, and requests canceled before generation starts deserve a clear policy, ideally with automatic credits where appropriate.

Give customers a usage history with clip length, model, date, and status. Support conversations become much easier when both sides can inspect the same record instead of debating what happened inside a vague project allowance.

Choosing a model customers can understand

The best billing model is one customers can predict before they click Generate. Start with the unit your product can reliably measure, then present it in the language your customers already use. For video features, generated seconds are usually that unit.

A practical setup for the commerce platform might include a monthly allowance of generated seconds, clear duration choices for each clip, and an additional charge or credit deduction when merchants exceed their included usage. Campaign bundles can sit on top of that structure when the platform limits duration and attempts.

Keep model choices understandable as well. If one model has a higher generation cost, show that it uses more credits per second or belongs to a higher plan. Customers can then choose between speed, quality, and budget with the relevant information in front of them.

Per-second billing gives variable video workloads a transparent foundation. Project pricing adds convenience when the workflow is tightly bounded. Teams that combine those ideas deliberately can offer flexible creation tools while keeping revenue, margin, and customer expectations aligned.