Below: what you need, how the browser workflow runs from start to finish, which devices work best and how your files are handled.
How OmniHuman 2 online works
The whole process takes a couple of minutes of your time; the render itself runs in the background.
- 1Add a photoDrag in a portrait or tap to pick one from your camera roll.
- 2Add a voiceUpload an MP3 or WAV, or press record and speak into your microphone.
- 3Direct it (optional)Describe gestures or a camera move in a sentence.
- 4Preview and renderWe validate your files for free, then render once you choose a plan.
- 5DownloadGet a 720p MP4 without watermark, or find it later in My videos.
Works on desktop and mobile
The studio runs in current versions of Chrome, Safari, Edge and Firefox. On phones you can shoot a selfie and record your voice without leaving the page. Recording uses your browser’s microphone permission; if you block it, uploading a file works just as well.
Because the heavy lifting happens on our servers, an older laptop or a mid-range phone produces exactly the same video as a high-end workstation.
What files you can use
Photos: JPG, PNG, WEBP and other common formats up to 30 MB, at least 300 pixels per side. Audio: MP3 or WAV between 4 seconds and 5 minutes, up to 15 MB. Browser recordings are converted automatically to a format the model accepts.
Privacy and storage
Uploads go to private storage that only you can access through time-limited links. Your files are sent to the model provider only when you start a paid render, and projects you never render are deleted after 30 days. We don’t use your photos or voice to train models.
Tips for the best browser results
When recording in the browser, use headphones so the microphone doesn’t pick up playback, and hold the phone or laptop mic about a hand’s width from your mouth. Speak the script once as a warm-up; the second take almost always sounds more natural.
For photos taken on a phone, turn off portrait-mode blur and beauty filters. Strong background blur confuses edges around the hair and shoulders, and heavy smoothing makes skin look plastic once it moves.
Your library and sharing
Every render is saved to My videos, where you can replay, download or start a new version from the same project. Downloads are standard MP4 files, so they upload directly to TikTok, Instagram, YouTube, LinkedIn or any editor without conversion. Links we generate for playback expire after a short time, which keeps private videos private.
FAQ
Do I need to download any software?
No. Everything runs in your web browser and on our servers.
Can I use it on my phone?
Yes. Upload from your camera roll and record your voice directly in mobile browsers.
Do I need an account?
Not to upload and preview. You sign in with Google to save your project and render.
How fast is it?
Short clips usually finish in a few minutes. You can leave the page while it renders.
Are my uploads private?
Yes. They are stored privately and used only to render your videos.
How much does it cost to use OmniHuman 2 online?
Previews are free. Rendering uses 10 credits per second on the Pro plan from $19.99 per week.
Which browsers work best?
Current versions of Chrome, Safari, Edge and Firefox. Recording needs microphone permission; uploading works everywhere.
Can I leave the page during rendering?
Yes. Renders run on our servers and the finished video appears in My videos.
Is there a file size limit for uploads?
Photos can be up to 30 MB and audio up to 15 MB. Larger recordings can be trimmed or exported as MP3 to fit.