Type vozo.ai/face-swap into a browser and you arrive somewhere unexpected: the homepage, which wants to translate and dub your videos in 160-plus languages. No face swap widget, no upload slots, nothing. The tool itself still exists. It moved into the Vozo web app, under a tab called Labs, behind a login, a consent dialog, and a two-run free allowance.
That relocation explains most of the confusion around this tool in 2026. The old landing page was still live in an Internet Archive capture from October 2025; today the same address bounces straight to the homepage. The pricing page never mentions the feature at all. The one place that states its cost is a documentation table three clicks deep. This piece walks the route that works, collects the numbers Vozo scattered across its docs, changelog, and client code, and sets a metered per-run option next to it for the jobs a subscription pool fits badly.
Key Takeaways
- Vozo's face swap is a video tool in the web app's Labs section. It replaces the face in an uploaded clip with a face from a photo; it does not process still images.
- Free accounts get 2 runs. After that, the tool requires a paid plan, and those start at $29 a month for 150 points.
- A swap draws 5 points plus 5 per minute of uploaded video, with the app's own cost calculator metering fractions in 12-second blocks. A 1-minute clip costs 10 points, about $1.93 at the Creator plan's monthly rate.
- The uploader caps clips at 3 minutes, 1 GB, and 1080p. None of those caps appear on the docs page; they surface in the app's tooltips and client script.
- Videos with several faces get every detected face replaced. Vozo's own advice is to upload single-face footage.
- Finished swaps are stored for 30 days. Vozo sells no face swap API, and its CLI skills omit the feature.
- For short clips, Atlas Cloud runs a dedicated face-swap-video endpoint with no plan attached: listed from $0.2 per second, with playground runs quoting $1.098 at 720P and $1.2996 at 1080P on July 30, 2026.
Where Vozo Face Swap Actually Lives
The working route is short once you know it exists. Sign in at app.vozo.ai, open the dashboard, and click Labs in the navigation. At the end of July 2026 that board holds two cards: Vozo Skills, an agent integration in beta, and Face Swap, described as "Replace faces in your videos seamlessly while retaining natural expressions and movements."
Click the card and a dialog titled Content Editing Consent appears before the tool does. Its text is short enough to read in full: "By using this feature, you confirm that you have legal rights to edit this content and will not use AI tools for illegal, non-consensual, or harmful purposes. Violation may result in account termination." Agreeing is remembered in your browser, so the dialog shows once. The app's client code also blocks the tool on phones for free users, with a notice that says the feature needs a desktop.
Redirect behaviour, Labs layout and dialog wording observed on vozo.ai and in its client code, July 30, 2026.
The Labs label is worth taking literally. Vozo's help page for the feature files it under "Lab features," the label the company uses for experiments outside its core translation lineup. That status shows in how little the rest of the site acknowledges the tool: the pricing page lists seven AI tools by name and face swap is not among them, the roadmap page does not mention it, and the changelog has touched it exactly once since mid-2025, to shorten how long results are stored. The tool works. It is also clearly not the product Vozo is building its business around.
How to Face Swap a Video in Vozo, Step by Step
The tool itself is three numbered steps stacked on one page.
- Upload the original video. This is the footage whose face will be replaced. The uploader accepts mp4, mov, webm, avi, and wmv files up to 1 GB, at up to 1080p, and up to 3 minutes long.
- Upload the target face photo. One photo in png, jpeg, jpg, or webp, up to 100 MB, larger than 200×200 pixels. This is the identity that ends up in the footage.
- Press Swap Face. The app shows the point cost before the run starts. During busy hours the job may sit in a queue first, which the docs warn about. If you cancel a queued task, the points come back.
Those caps deserve a second look, because the documentation page states none of them. Every limit above comes from the tooltips on the upload buttons and from the limits object in the app's client script, which sets video duration at 0 to 180 seconds and resolution at 1920 by 1080. If your clip runs 3 minutes and 5 seconds, the uploader refuses it, and nothing on the help page will have warned you.
The multi-face behaviour is the other planning fact. Vozo's docs say it plainly: "For videos with multiple faces, all detected faces will be replaced." There is no face picker. A two-person interview becomes two copies of the same person. The workaround Vozo suggests is choosing footage with one face in frame, which is honest advice and also a real constraint on what this tool can do.
Results land in a Generated Videos strip under the form, playable in place, with a download menu per clip. They stay there 30 days.
A Metered Vozo Face Swap Alternative for Short Clips
The Labs tool wants a subscription before your third swap. Whether that trade makes sense depends mostly on the shape of your footage, and there is a second route worth knowing before you commit either way. Atlas Cloud runs a dedicated face-swap-video endpoint in its tools lineup, and it meters per run instead of per plan.
The mechanics match Vozo's: a source face photo, a target video, one button. The endpoint takes MP4 or MOV clips of two to ten seconds and outputs at 720P or 1080P, with motion, timing, camera movement, and background preserved while the face changes. The catalog lists it from $0.2 per second. In the Face Swap video playground, a run quoted $1.098 at 720P and $1.2996 at the 1080P default on July 30, 2026, with no discount active on the endpoint that day. A sibling endpoint does the same job on still photos from $0.09 per image, which covers the photo case Vozo's video-only tool leaves out.
Vozo column from its help page, points and subscription docs, and client code; Atlas Cloud column from the face-swap-video and face-swap-image model pages. All retrieved July 30, 2026.
One row goes decisively to Vozo, and it is the row to check first: clip length. Vozo takes uploads up to 3 minutes, the metered endpoint takes 2 to 10 seconds, so a talking-head video is work only Vozo can do in a single pass. Almost everything else on the table runs the other way. A meme, an ad variant, or a social clip fits both routes, and then the question is what you are buying: a monthly pool with an expiry date and a concurrency cap, or a per-run price that is printed on the button, carries no balance, and answers a REST call when the job stops being a one-off.
Method 1: Run a Video Face Swap in the Atlas Cloud Playground
No plan or code required, just an account with credit. The form has three fields: Source Face for a clear, front-facing photo of the identity going in, Target Video for the clip being changed, and an Output Resolution toggle between 720P and 1080P. The price on the Run button updates when you switch resolution, so the cost is visible before anything is charged. Output arrives as a playable result you can download directly.
Method 2: Automate Video Face Swaps Through the Atlas Cloud API
The same endpoint takes scripted traffic, which is the piece Vozo does not sell at any tier. Three steps.
Step 1: Create an API key in the Atlas Cloud console dashboard and copy it.
Step 2: Read the API docs for authentication and the shared video endpoint.
Step 3: Send the request. Face swap rides the platform's standard generateVideo endpoint, with the model string doing the routing:
Bash1curl -X POST https://api.atlascloud.ai/api/v1/model/generateVideo \ 2 -H "Authorization: Bearer $ATLASCLOUD_API_KEY" \ 3 -H "Content-Type: application/json" \ 4 -d '{ 5 "model": "atlascloud/face-swap-video", 6 "image": "https://your-cdn.com/source-face.jpg", 7 "video": "https://your-cdn.com/target-clip.mp4", 8 "resolution": "720P" 9 }'
Generation is asynchronous. The response returns a prediction ID that you poll until the job completes:
Bash1curl https://api.atlascloud.ai/api/v1/model/prediction/<prediction_id> \ 2 -H "Authorization: Bearer $ATLASCLOUD_API_KEY"
The finished clip comes back as a URL in the outputs array. Because every model on the platform shares this endpoint pattern and key, swapping the model string to the image endpoint, or to any video generator in the catalog, reuses the same code. In practice that turns fifty product clips with fifty faces into a short script rather than an afternoon of manual uploads.
What a Vozo Face Swap Costs in Points
Vozo prices the swap in one sentence on its points usage page: 5 points base plus 5 points per minute of uploaded video, with a single worked example, 10 points for a 1-minute video. The rounding rule lives in the app instead. The cost tooltip next to the Swap Face button adds a second example, 8 points for a 30-second clip, "rounded up," and the client script behind it computes the charge in 12-second blocks, one point each. A 61-second video therefore costs 11 points, not the 15 a whole-minute reading would suggest.
Vozo figures from its points docs, the app's cost tooltip and its pricing page; Atlas Cloud from the face-swap-video playground at 720P. Both July 30, 2026. Dollar figures for Vozo derived at $29 per 150 points.
The bottom panel is the part that decides most budgets. Vozo's cheapest plan is a fixed $29 whether you run three swaps or twenty-five, so the per-swap cost only falls to something competitive if you spend the whole allowance on faces. Run three short clips in a month and each one has effectively cost you $9.67. The metered endpoint charges $3.29 for the same three. The two lines converge near the top of Creator's capacity, at which point Vozo is the better rate, and only if nothing else touched the pool.
Points are sold only inside plans, so the dollar cost depends on which one funds the pool. The pricing page sells Creator at $29 a month for 150 points, Studio at $99 for 600, Studio XL at $249 for 1,500, and Studio XXL at $649 for 4,000, with yearly billing cutting the monthly figure to $22, $75, $188, and $500 and issuing the year's points as one allocation, valid for twelve months instead of the rolling two. On Creator's monthly rate a point runs about 19 cents, which makes a 1-minute swap roughly $1.93 and fills the month with at most 15 of them, assuming nothing else touches the pool. Dubbing, lip sync, and every other point-metered tool draw from the same 150.
The free tier is smaller than it looks. New accounts get 20 gift points, valid for 7 days, and the face swap tool separately enforces its own cap of 2 runs per free account. The two limits meet neatly, since a pair of 1-minute swaps costs exactly 20 points. After that, the upgrade dialog takes over. The client strings also sketch a face-swap-specific offer, a 50-point pack pitched at new users and valid 30 days; treat that as something the app may show you rather than a published price, since no public page lists it.
One more caveat carries over from the plan cards: watermark removal is listed as a paid-plan feature. The face swap docs never say whether Labs output is marked on the free tier, so budget for the possibility that your two free runs are watermarked.
Vozo Face Swap Rules That Sit Outside the Tool
Nothing on the swap page itself mentions retention, yet three separate time limits apply to your results, each documented somewhere else on the site.
Compiled from Vozo's help page, points rules, subscriptions page, changelog, and client strings, July 30, 2026.
The 30-day storage window is the one that catches people, because it used to be longer. Vozo's changelog entry for July 24, 2025 states it in one line: "Face Swap videos are now stored for 30 days only." The tool's own interface repeats it above the results strip. Download finished swaps as soon as they render, because the dashboard copy is the only one that exists.
The other two limits belong to the subscription machinery. Vozo's billing page says stored projects are deleted 90 days after a subscription expires, and that unused points freeze the moment a plan lapses, usable again only if you resubscribe. Monthly point allocations expire two months after they arrive. None of this is unusual for a subscription product. It just never appears next to the swap button, so it is worth writing into your own production schedule.
Automation has no rules to list because it has no route. The API suite Vozo sells covers translation, dubbing, visual translation, lip sync, and talking photos; face swap is absent. The Vozo Skills integration announced in July 2026 connects AI agents to dubbing, subtitles, visual translate, and lip sync, and skips face swap as well. If a workflow needs face swaps on a schedule or at volume, this tool cannot be wired into it, which is the clearest single line between Vozo and the metered endpoint above.
Frequently Asked Questions
Is Vozo AI video face swap free?
Twice. Free accounts can run the face swap tool 2 times, funded by the 20 gift points that come with registration and expire after 7 days. From the third swap onward you need a paid plan, and those start at $29 a month, or $22 a month billed yearly.
Can Vozo face swap a photo instead of a video?
No. The docs state the feature "currently supports videos only," and the second upload slot takes a photo only as the face source. For stills, a dedicated image endpoint on Atlas Cloud swaps a face onto a photo from $0.09 per image, with the playground quoting $0.096 per run.
Does Vozo have a face swap API?
No. Vozo's published API covers translate and dub, visual translate, lip sync, and talking photo, and its CLI skills cover a similar list. Face swap appears in neither. Scripted face swaps need a different provider; Atlas Cloud's face-swap-video endpoint answers a standard REST call at the same endpoint pattern as its other video models.
How long does Vozo keep face swapped videos?
30 days, a limit set in July 2025 and repeated in the tool's interface. Separately, if your subscription ends, Vozo deletes stored projects 90 days later. Download anything you want to keep as soon as it renders.
Why does a Vozo face swap replace everyone in the frame?
Because the tool has no face selector. Vozo's docs warn that all detected faces get replaced and recommend uploading videos with a single face in frame. For group footage, that means cropping or choosing different source material before the swap, not after.
What video formats does Vozo face swap accept?
The uploader takes mp4, mov, webm, avi, and wmv, up to 1 GB per file, up to 1080p, and up to 3 minutes. The face photo can be png, jpeg, jpg, or webp up to 100 MB, larger than 200×200 pixels. These caps appear in the app's upload tooltips rather than in the documentation.
Conclusion
Vozo's face swap is a working tool that its own maker still files under experiments. Swap quality is yours to judge inside the two free runs, but the logistics are already on record: video only, one face per frame works, three minutes maximum, results gone in 30 days, and a subscription standing between you and the third swap. For someone already paying for Vozo's dubbing or lip sync, the marginal cost of a swap is a few points from a pool they own anyway, and the 3-minute ceiling comfortably covers talking-head work no short-clip endpoint will take in one pass.
The calculation flips when face swapping is the whole job. A plan bought for one Labs feature means paying $29 a month for a tool the pricing page does not advertise, with no API to grow into. Short clips, scripted volume, or once-in-a-while swaps fit a per-run price better: the meter reads before you commit, the code path exists, and nothing expires at the end of the month.
