Playbook
How Magic Hour AI gets its Google traffic
Two face-swap searches bring Magic Hour AI an estimated 91,453 monthly US organic visits.
- Estimated US organic visits/month
- 203,411
- Estimated ranking keywords
- 15,743
- Estimated non-brand top-20 keywords
- 5,993
- Estimated AI Overview citation searches
- 2,567
- Estimated referring domains
- 10,150
- Domain rank
- 60
The playbook
Magic Hour wins searches for specific creative tasks: just “face swap” and “ai face swap” contribute an estimated 91,453 monthly US visits. Working product pages earn the biggest rankings and attract links; brand terms supply only an estimated 4.0% of the top 100 keywords’ traffic. The first move to copy is one useful, free task page that lets visitors get a result before signing up.
Scorecard
These are search estimates and backlink-index measurements, not Magic Hour’s internal analytics.
| Metric | Finding |
|---|---|
| Estimated monthly US organic traffic | 203,411 |
| Estimated ranking keywords | 15,743 |
| Estimated brand share of top 100 keywords’ traffic | 4.0%: 6,950 of 174,966 estimated monthly visits |
| Estimated non-brand keywords in the top 20 | 5,993 |
| Estimated searches with an AI Overview citation | 2,567, including brand searches |
| Referring domains in the backlink index | 10,150 |
| Referring domains: 12-month change | unknown for an exact year; the available monthly series rises from 6,221 on 2025-11-30 to 9,874 on 2026-10-31, up 3,653 (+58.7%) |
| Domain rank | 60, a provider score |
Brand matching includes “magic hour,” “magichour” and “magic-hour.” The brand percentage describes the top 100 keyword sample, not all traffic.
Traffic engines
Shares below are estimated shares of non-brand top-20 keyword rankings, not visits. How pages are made is inferred from their visible structure; their actual publishing process is unknown.
| Section | Page type and how it is made | Estimated non-brand top-20 keywords | Estimated share | Top page | Top keyword |
|---|---|---|---|---|---|
| /products/ | Working generators and editors; shared layout with task-specific copy and controls | 3,764 | 62.8% | Face swap | face swap |
| /tools/ | Narrow task pages; appears to use one template filled with preset, example and task data | 1,458 | 24.3% | Celebrity look-alike finder | celebrity look alike |
| /templates/ | Individual reusable clips; appears to use one template filled with asset data and supporting copy | 336 | 5.6% | Guy Fieri facial expression | guy fieri ai video |
| /blog/ | Authored reviews and explainers; appears individually written, with a shared article layout | 333 | 5.6% | Pixlr review | pixlr face swap |
| Rest of site | Other paths and subdomains; production method unknown | 102 | 1.7% | unknown | unknown |
The leading pages show how search visitors enter the product:
| Page inspected | Search need | Visible path into the product | Publication date shown |
|---|---|---|---|
| Face swap | Replace a face in a photo | Two upload fields and a swap button appear first; links lead to video, GIF and multiple-face modes | unknown |
| Celebrity look-alike finder | Find celebrity resemblances | Upload a portrait and generate suggested matches; links lead to the full image editor and its models | unknown |
| Guy Fieri template | Make a specific celebrity reaction clip | A face upload and Generate Video button sit beside the ready-made asset; related links lead to face swap and other products | unknown |
| Pixlr review | Understand a competing face-swap option | A dated, named-author review explains inputs and limitations, with a Try Magic Hour button | September 22, 2026 |
Top pages
These are 11 leading pages found in the 40 highest-traffic non-brand top-20 keyword rows. Traffic sums only those sampled rows per URL; it is not a complete page traffic estimate.
| URL | Section | Top keyword by estimated traffic | Estimated monthly searches | Estimated absolute position | Estimated monthly traffic across sampled rows |
|---|---|---|---|---|---|
| /products/face-swap | /products/ | face swap | 246,000 | 1 | 98,845 |
| /products/face-swap/video | /products/ | video face swap | 14,800 | 1 | 30,838 |
| /products/ai-video-generator | /products/ | free ai video generator | 90,500 | 3 | 18,278 |
| /products/image-to-video | /products/ | ai image to video | 22,200 | 3 | 4,111 |
| /tools/celebrity-look-alike-finder | /tools/ | celebrity look alike | 165,000 | 13 | 2,477 |
| /products/ai-gif-generator | /products/ | ai gif generator | 3,600 | 2 | 1,126 |
| /products/body-swap | /products/ | body swap | 12,100 | 3 | 758 |
| /products/lip-sync | /products/ | lip sync | 18,100 | 3 | 376 |
| /products/multiple-face-swap | /products/ | multiple face swap | 2,400 | 2 | 370 |
| /blog/pixlr-face-swap-review | /blog/ | pixlr face swap | 22,200 | 9 | 280 |
| /tools/twerk-generator | /tools/ | ai twerk generator free unlimited | 480 | 1 | 276 |
Search volumes are estimated 12-month averages. “Face swap” includes an estimated November 2025 spike of 823,000 searches versus 110,000 in September 2026; “celebrity look alike” includes an estimated September 2026 spike of 1,220,000 versus 74,000 in August.
Search patterns
The strongest patterns describe an action, an input format or a barrier the visitor wants removed.
| Pattern | Example keywords | Estimated monthly searches, respectively | Section answering it |
|---|---|---|---|
| [Action] / AI [action] | face swap; ai face swap | 246,000; 135,000 | /products/ |
| [Action] + [media format] | video face swap; face swap video | 14,800; 14,800 | /products/ |
| Free [task] + no sign up | free ai video generator; ai video generator free no sign up | 90,500; 5,400 | /products/ |
| [Input] to [output] | ai image to video; image to video | 22,200; 27,100 | /products/ |
| [Specific variation] + [action] | multiple face swap; ai body swap | 2,400; 1,000 | /products/ |
| Find a resemblance | celebrity look alike; who does this look like | 165,000; 12,100 | /tools/ |
| [Competitor] + [feature] | pixlr face swap | 22,200 | /blog/ |
| [Named subject] + AI video | guy fieri ai video | 40 | /templates/ |
The Pixlr average is also uneven: estimated monthly searches range from 1,900 to 74,000 in the series. These averages should not be treated as steady monthly demand.
Link magnets
Of 50 domains sampled in descending domain-rank order, 35 remain after excluding user profiles, redirects, search pages and a canonical reference. Counts describe retained source domains, not all links or independent editorial endorsements.
| Page type | What it is | Strongest sampled domains pointing to it | Examples |
|---|---|---|---|
| Homepage | The overall product and company | 19 | Figma’s design-tool guide, Jotform’s headshot guide, Wix article |
| Product pages | Usable face-swap, image and video features | 10 | Mango Animate’s face-swap roundup, Featurebase’s growth article, Tuts+ image-generator guide |
| Blog articles | Model pricing, release tracking, comparisons and statistics | 5 | Sora pricing reference, Video-model tracker reference, AI-art statistics reference |
| API documentation | Instructions for uploading assets | 1 | Magic Hour’s developer article |
Products are the clearest repeated link destination beyond the homepage. The sample also includes self-published pages, forums and press distribution, so a strong host-domain score alone does not prove a strong endorsement.
Moves to copy
| Move | Evidence at Magic Hour | How a smaller startup can do it | First page to build |
|---|---|---|---|
| Make one core task usable on its search page | The face-swap page contributes an estimated 98,845 monthly visits across its sampled keyword rows; its upload form appears before explanatory copy | Choose one task the product already performs well. Put the input, action and result on the landing page. Explain limits and show examples below it | A free page for the product’s most requested action, such as a CSV cleaner for a data startup |
| Give distinct formats their own workflows | The video face-swap URL contributes an estimated 30,838 monthly visits across sampled rows, separate from photo face swap | Split pages when the input and workflow genuinely differ. Link the variants together. Keep synonymous searches on the same useful page | A file-format variant of the proven task, such as cleaning an Excel file |
| Turn a broad feature into a specific preset | /tools/ holds an estimated 1,458 non-brand top-20 keywords; the celebrity finder leads directly into image editing | Package one common outcome with a preset and before-and-after examples. Let visitors use it without composing a complex prompt | A portrait-background cleanup preset for an image startup |
| Build something other publishers can recommend | Product pages receive links from 10 retained domains in the strongest-domain sample | Make the free feature reliable, easy to demonstrate and directly linkable. Publish clear inputs, outputs and usage limits | A standalone calculator, checker or converter tied to the paid product |
| Answer a competitor’s feature question with a useful review | The Pixlr review ranks at estimated position 9 for “pixlr face swap”; blog pages attract 5 retained sampled domains | Write one sourced comparison for a decision customers already make. State where each option fits and date the review | “[Competitor feature] review: inputs, limits and alternatives” |
What won’t copy
| Advantage or limit | Why it matters |
|---|---|
| Existing brand demand | The top 100 keywords include an estimated 6,950 monthly visits from Magic Hour brand terms. A new startup cannot inherit that demand |
| An accumulated link profile | The backlink index records 10,150 referring domains, with sampled links first seen as early as 2023. Copying the page layout does not copy those relationships |
| A working creative product | Search visitors can upload files and generate results. The underlying software, output quality and cost of free usage are part of the offer |
| Large page counts | The earlier sitemap count is 19,766 pages, but /templates/ accounts for only an estimated 5.6% of non-brand top-20 keywords. Page count alone does not explain the leading rankings |
| Uneven demand and unknown conversion | Some large keyword averages contain sharp spikes. Revenue, signup conversion and profitability by landing page are unknown |
Method
Data from OpenSEO (DataForSEO estimates, US market), dated 2026-11-18, with English search settings and direct inspection of public Magic Hour pages.
The analysis uses the top 100 keywords for the brand split, 40 non-brand organic keyword rows for page totals, four section counts, AI Overview citation counts and 50 backlink rows sorted by source-domain rank. Page totals exclude duplicate rows from section checks. AI Overview counts include brand searches and are not visit estimates. Rankings and backlink records may have been observed before the report date.