AL-1
AI video calculator
What a film will cost to generate, before you open an account anywhere.
The number you budget is not the number you are charged
A generative model does not bill you for the film you deliver. It bills you for every second it produces, in the block sizes it sells — including the takes you discard, and the seconds inside a block you paid for and never used.
At a coverage ratio and a feature runtime, that difference stops being a rounding error. It is where AI video budgets go wrong, and it is the whole reason AL-1 exists.
Read from inside the accounts, so you can budget before you open one.
We opened accounts on 16 of the 25 platforms and read the prices from inside. A great many rates only exist behind a sign-in — the closest comparable tool labels some of its own figures “needs logged-in check”, which is a polite way of saying it could not look.
Your shot is 2.1 seconds. You are billed for 5.
Films are shots, and models bill in fixed clips — four, five, eight, ten seconds. AL-1 quantises your shot length to each model’s billable blocks and shows you the waste as its own line. No other calculator we found models it.
It says “I don’t know” instead of guessing low.
Where a route cannot be costed honestly, AL-1 withholds it and names what is missing. Where sources disagree, it keeps both and renders the spread. The purchase here is not saving money. It is not being wrong in front of a client or a financier.
The colours are a claim about evidence
AL-1 is a 1970s desk calculator, and it borrows the machine’s palette for one reason: on the hardware, a figure only glows when the tube is actually reading something. The same rule holds here. Colour tracks confidence, and nothing else.
- $2,160.00Verified
We read this rate from the seller, and we stand behind it.
- $2,160.00Estimate
We have not verified this one yet. Amber is a coverage map, not a warning: the gap is ours to close, not yours to absorb.
- $2,160.00No seller
It runs on your own power. There is no price to check, only your electricity and your wall clock.
Every figure carries the source it came from, the date it was read, and how far it is trusted — from the dataset through to the screen. A confidence tag is never quietly upgraded.
What it covers
Direct APIs, aggregators, subscription credit plans and local GPUs, in USD, EUR and PLN. For credit platforms it solves the plans and top-up packs, so the answer is not “you need 47,300 credits” but “Ultra for one month plus eight 5,000 packs”.
That is roughly five times the coverage of the nearest comparable tool, which publishes 14 models. The next publishes 15.
The card is append-only: a price change arrives as a new entry carrying the date it took effect, and the figure it replaced still says when it was true. AL-1 tells you how fresh the card is without being asked. Dataset 2.1.0, bundled with the app.
It collects nothing
No account. No analytics. No advertising. No tracking. No third-party SDKs. The entire rate card ships inside the app, so airplane mode still prices your film — nothing is fetched and nothing is uploaded. Your films, your saved scenarios and your settings stay on the device, because there is no server to send them to.
Status
Version 0.1.1, in TestFlight and not yet on the App Store. It will be a paid app at $1.99. English and Polish. The web build runs the same engine on the same dataset and is free to use.
Every figure AL-1 produces is an estimate, built from published and account-read prices. It is not a quotation and not a confirmed charge. Verify with the provider before committing spend.
Provider and model names are used only to identify whose service is being costed. No affiliation with, or endorsement by, any named provider is implied.