How ScaleStack ranks AI video makers
We score each tool from public vendor pages only, on six criteria, using 0, 3 or 5 points per criterion. Nothing else enters the calculation. This page is the full method for the AI video makers category (draft 1, 2026-10-16).
What "Desk-researched" means
- We read public pages: the official pricing page, product pages, documentation, terms, changelog and public marketplace listings.
- We write down what those pages say on the date shown. We do not create accounts or run the tools. Every page carries the label Desk-researched. We will not use the word "Tested" until a person or agent has run a real job through a tool and kept the notes.
- Claims about output quality and reliability come from published specs and documents. Each page says these are the vendor's published claims, not our results.
- Every record shows last checked (the date we read the page) and the source (the URL). The last checked date changes only when we re-read the source.
Category
AI video makers: software that makes a finished video from a script, prompt, images or uploaded footage. Pure editors, screen recorders and done-for-you video services are out of scope unless we say otherwise on the ranking page.
Scoring math
- Each criterion has a weight. The weights add up to 100.
- For each criterion a tool scores 0, 3 or 5. We pick the highest anchor the public evidence fully meets. If the evidence does not fully meet the 3 anchor, the score is 0. There are no other scores.
- Criterion points = score ÷ 5 × weight. Tool total = sum of criterion points, out of 100.
- Rank by total, highest first, using unrounded totals. Totals display to one decimal place. Tools with equal totals share a rank and are listed in alphabetical order.
- Nothing else enters the calculation.
Criteria, weights and anchors
| Criterion | Weight | 0 points | 3 points | 5 points | Where to check |
|---|---|---|---|---|---|
| Output quality | 25 | No page publishes max export resolution and max video length for any paid plan | Pricing or docs page states max export resolution (1080p or higher) and max video length for at least one named paid plan | Everything in 3, and the pricing or docs page states 4K export on a named plan or a single-video length of 10 minutes or more on a named plan, and lists supported export formats or aspect ratios | Pricing page plan table; product docs or help center "export" article |
| Price and value | 20 | No price published for any paid plan (contact sales only) | A paid plan's price is published with its billing period and the limit it buys (minutes, credits or videos per month) | Everything in 3, and the price per published minute of video on the lowest paid monthly plan is the lowest in the ranked set, or within 10% of the lowest. Credit-based plans only count if a page states how many credits one minute costs; otherwise the tool scores 3 | Pricing page; credits explainer in help center |
| Ease of use | 15 | No page describes the onboarding or editor | A help article or product page describes an editor and a template library (the template gallery is public) | Everything in 3, and a help article describes a prompt-to-video or script-to-video flow that produces an exportable video without a manual editing step | Product page; help center "getting started" and "create a video" articles |
| Free tier | 15 | No free plan and no free trial published | A free plan or free trial is published, with a watermark on exports or tight caps on length or credits | A free plan is published and its page or terms say exports have no watermark | Pricing page; plan comparison table; terms of service |
| Fit for the use case | 15 | The pages document 0 or 1 of the 4 fit features below | The pages document 2 or 3 of the 4 fit features | The pages document all 4 fit features, and the product page or docs name small business, local business or ad creation as an intended use | Product page; features page; terms of service (commercial use clause) |
| Update pace | 10 | No dated product or changelog entry in the 12 months before the check date | A dated entry in the 6 months before the check date | At least 2 dated entries in the 90 days before the check date | Changelog, release notes, product blog with dates, help center "what's new" |
Weight total: 25 + 20 + 15 + 15 + 15 + 10 = 100.
The four fit features
The use case for "Fit" is turning phone footage or a script into short social or ad videos for a local service business. The four features are:
- Upload your own video or photos
- Automatic captions
- Vertical 9:16 export
- Terms that allow commercial use on the plan we point to
The UNVERIFIED rule
- If a criterion cannot be checked from a public page on the check date, it scores 0 and the page shows UNVERIFIED next to it.
- A page behind a login, a form, a "contact sales" wall or a call booking does not count as public.
- If a source disappears or changes and we have not re-read it, the field goes back to UNVERIFIED. We do not keep an old claim.
- Tool pages count the UNVERIFIED fields. Ranking pages show how many criteria are checked for each tool.
Money never changes rank
- Rank is computed only from the rubric scores.
- Rankings are never ordered by payout, price paid to us, or who asked us to list a tool.
- Commission data is never used for ranking. The scoring script reads only the public tool data and the score sheet.
- This site has no affiliate or sponsored links at launch. Links to vendor sites are plain links with no tracking.
What we do not score
- Our own results or opinions of output. We have not generated videos, so the label stays Desk-researched.
- Sample video galleries and customer success claims.
- Payments, sponsorships, or whether a vendor has contacted us.
- Third-party star ratings. We may cite counts with source and date, unscored.
- Brand size, funding or social following.
- Legal or compliance advice. We may note what a vendor's terms say about commercial use, likeness and AI disclosure, unscored.
Source rules
Public pages only: pricing page, help center, terms, changelog. We record the URL, the date printed on the page (or "no date shown") and the date we read it.
Draft status
A ranking page is marked DRAFT - not ready to publish: scores incomplete until every criterion for every tool has been checked against a source page. See the AI video makers ranking for the current state.
How we pick use cases
The home page is a feed of real AI use cases for businesses. These rules apply to every card. They are separate from the scoring above, and use cases are never scored or ranked.
- A public source is required. Each card links to one public page (a company blog, case study, documentation, news article or public post) that we opened and read. Paywalled and login-only pages do not count. The card shows who reported it and the date printed on the page, or "no date shown".
- We summarise in our own words. Titles and summaries are written by us, not copied from the source. If a short quote is ever needed, it is under 15 words and in quotation marks.
- No results claims shown as fact. Income, revenue, return on investment and similar figures are left out. If a source's figure ever appears, the card says "the source reports" and names the source. Many sources are vendors' own marketing pages, so each card says "reported by" the vendor. A vendor's account is not independent proof.
- Tools are listed as the source names them. A tool links to its ScaleStack page only if we have one. Other tools show as plain names.
- No affiliate links or tracking. Source links are plain links.
- Uncertain fields are marked UNVERIFIED. A card without a source we read does not go on the page.
Corrections
- Report an error at hello@scalestacktech.co.
- We fix confirmed errors and show the date of the change on the page.
- If a correction changes a score, the rank is recalculated from the rubric. A vendor cannot ask us to change a score without a public page that supports it.