Try your first 6-second video free. No card required.Explore plans
OmniHuman 2.0
Image + audio. A whole new performance.

OmniHuman 2.0 AI Video Generator

Give your image a voice. Turn a portrait and audio into expressive video, with natural lip sync, gestures and movement.

Create your videoPrivate by default
Portrait image
Click to upload or drag and dropJPG, PNG or WebP · up to 10 MB and 4 megapixels

Scan to upload a photo

Scan this QR code with your phone, choose a photo, and it will appear here automatically.

Creating a secure upload link…

Keep this window open. Anyone with this QR code can send one photo to this window.

Audio track
Click to upload or drag and dropMP3 or WAV · 4–300 seconds · up to 15 MB
0/300
ModelOmniHuman 2.0
720p10 cr / s

Preview your cost without signing in. 10 credits = 1 billed second.

First video free, up to 6 seconds after email verification. No card required. Dollar equivalent uses Starter monthly (about $0.82/s). Other plans have different rates.

Private by default↺ Failed generations return their credits

First video free, up to 6 seconds after email verification. No card required. Your creations stay private.

OmniHuman 2.0 image to video generator: input and credit help

Use the fifteen-minute quote window

A validated quote remains available for fifteen minutes. It covers the uploaded files associated with that task and shows their measured duration and credit cost. Review it while those inputs are still the version you want to use. If you decide to edit the recording or replace the portrait, return to input validation and obtain a new quote for the changed sources. Once the original quote has expired, the site asks you to upload and check the files again. Expiry protects the relationship between the files, their validated duration, and the confirmation you see. It does not mean that a previously submitted generation has failed. For a task already submitted, check the library status instead of treating the expiry of an earlier preview as a reason to submit again.

Separate source editing from performance direction

Use an audio editor when you need to remove silence, replace a spoken sentence, adjust the recording level, or shorten the overall message. Use an image editor when the portrait needs a tighter crop or a smaller pixel size. Those changes belong in the source files before validation. Inside this workspace, the optional direction concerns the visible performance, such as expression, gestures, posture, and camera movement. It is not a replacement for changing the source recording. Preview the exported files yourself before submitting them: an image that looks acceptable as a small thumbnail may hide an obstructed mouth, while a recording that starts cleanly may still contain an unwanted ending. Preparing the inputs first makes the quote easier to assess and the generated result easier to review.

Prepare both required source files

This workspace requires one portrait and one audio recording. If you currently have only a script, record it or create an audio file in a separate tool, then listen to the exported version. An optional direction can guide the performance, but it does not provide text-to-speech or replace either required file. Choose a portrait representing the speaker or character you intend to animate. Confirm your permission to use both sources, including permission from identifiable people where applicable. Visitors can explore the workspace and type a direction before registering. Starting generation requires a signed-in account with sufficient purchased credits. Creating an account is free and starts with a zero credit balance.

Check the actual image dimensions and format

Upload a still JPEG, PNG, or WebP no larger than 10 MB. Width and height must each be between 300 and 6000 pixels, the total area must not exceed four megapixels, and the width-to-height ratio must fall between 0.4 and 2.5. For instance, a 1200 by 1600 portrait satisfies these dimensions. A narrow panoramic crop can fail the ratio check even when the file is small. Export a resized copy when necessary. Renaming an unsupported file with a PNG extension does not convert its format. Animated containers are rejected, and validation decodes the complete image so an incomplete file with a plausible header cannot pass as a valid portrait.

Listen to the recording before uploading

Supported audio is MP3 or uncompressed PCM or floating-point WAV, up to 15 MB and between 4 and 300 seconds long. Remove an accidental count-in, long silence, or unwanted ending before upload. The recording determines video duration, so editing it first can reduce both the performance length and the quote. Changing an extension does not convert another audio format; export a fresh supported file instead. The site calculates duration from encoded MP3 frames or WAV samples rather than trusting an editable duration tag. If audio validation fails, check the exported format and length in your editor. A prompt asking for a shorter video does not trim the uploaded audio.

Write a focused direction within 300 characters

Describe a few observable choices: steady eye contact, relaxed shoulders, subtle hand gestures, or a fixed camera. A musical performance might call for a more expressive face or movement that follows the recording. Avoid conflicting requests such as a completely still subject and energetic whole-body motion in the same direction. The prompt provides guidance rather than exact keyframes, so you should inspect the result before relying on it. Leave this field empty if the source image and recording already communicate what you want. When comparing versions, change one aspect of the direction at a time and keep the same sources. Each separately confirmed new generation receives its own credit charge.

Read the verified quote before confirming

Select Preview cost without signing in to validate the inputs. The calculation rounds audio upward to the next whole second and multiplies it by 10 credits. A 4.2-second recording costs 50 credits, a 60-second recording costs 600 credits, and a 300-second recording costs 3000 credits. These are calculation examples rather than measured output samples. Compare the quote with your available balance. Validation and final generation are separate steps; confirm only when the files, direction, and cost are correct. If a quote expires, recheck your files to obtain a fresh one. Verified accounts receive 60 trial credits once for a first 4–6 second video. A subscription or credit pack adds balance for longer videos.

Follow the original task after submission

The confirmation deducts the quoted credits and sends the task for processing. Completion time varies, so there is no fixed delivery-time promise. Your library records the attempt and its status. If the connection drops, return to that existing attempt before submitting a separate copy. Repeated confirmation of the same task is handled as one submission, while starting a fresh task creates a separate generation. A provider-confirmed failure automatically returns the credits used for that attempt once. An uncertain response can leave a task awaiting confirmation while processing continues. The site keeps that attempt for reconciliation rather than immediately refunding and repeating a request that may already be running.

Review, download, and manage your balance

After the finished file has been saved, your private library provides playback and download. The output is 720p, with duration following the audio. Check the face, mouth movement, hands, and start and end of the clip before sharing. Generative motion can vary between attempts; a new creative version needs a fresh quote. Monthly subscription allowances expire at the end of their allowance month. Annual plans are billed yearly but deliver allowances monthly, rather than unlocking twelve allocations immediately. Separately purchased packs never expire. Expiring credits are spent first when both types are available. A returned credit still follows the expiry policy of the original allowance used for the attempt.

Troubleshoot using the file rules and task status

For a rejected portrait, check its real format, animation status, dimensions, ratio, and pixel count. For audio, export a fresh MP3 or supported WAV and verify the 4-to-300-second range. For an insufficient balance, review the available plans or packs and return after purchase. For an unresolved attempt, check whether it is processing, awaiting provider confirmation, or saving its output. If you need support, provide the task identifier and a brief description; never send a password or API credential. Your files are private by default, but the generation provider still processes them to perform your request. Review the privacy and acceptable-use policies before submitting personal or sensitive material.

Model input requirements: KIE OmniHuman 2.0 documentation. Explore subscriptions and credit packs

FAQ

A few things to know.

What do I need to make a video?

One portrait image and one MP3 or WAV audio track. You can also add an optional prompt to guide the scene, movement or camera.

Can I generate a video from text alone?

This model requires an image and audio. Your optional text prompt provides creative direction; it does not replace either input.

How are credits calculated?

Select your image and audio to preview the cost without signing in. Audio duration rounds up to whole seconds, at 10 credits per second. A 30-second clip uses 300 credits. Each plan shows its dollar equivalent per second. Sign in only when you are ready to confirm generation.

Do I receive free credits when I register?

Yes. Verify your email to receive 60 credits once, enough for one 4–6 second video. No card is required. Longer videos require purchased credits. Repeated sign-ins do not reset the trial, and confirmed failed generations return the credits.

Who can see my videos?

Your videos and uploads are private by default. Sign in to your account to view and download your work.

What happens if generation fails?

When the provider confirms a failed generation, the credits used for that task are returned automatically.

How do annual subscriptions work?

You pay for the year once and receive a new credit allowance every month. Each monthly allowance expires at the end of its allowance period.