Do I need to run ComfyUI myself?

No. The site runs its hosted workflow. You need a Google account and available site credits, not a local GPU or personal API key.

Will every result look like the homepage film?

No. The homepage fantasy clip is a selected result with separate upscaling and audio. Public generation uses its standard output settings.

Can I precisely lock a camera path or character?

There is no public camera-path editor or cross-task character lock. Clear references and simple motion can help, but exact control is not guaranteed.

FreeVideo AI: credits, privacy and generation limitations

Accounts, credits and queues

Google sign-in is required. Eligible new accounts receive one-time 60-credit welcome grants valid for 30 days. Text tasks cost 6; image and first-and-last-frame tasks cost 8. Only one task per account can remain active. Processing time varies with queue and GPU availability; there is no fixed completion-time promise.

Failures and retries

A confirmed failed task can reverse its linked credit charge. A missing history entry or a network interruption does not establish failure. A successful but unsatisfactory video is not automatically eligible for a free retry. Save the task ID and avoid repeated submissions while status is unresolved.

Uploads and generated files

Prompts and references are sent to the configured ComfyUI generation service. Supported reference files are PNG, JPEG and WebP, up to 10 MB each. Signed-in owners and authorized administrators can access private task results. The service does not promise a uniform automatic deletion or permanent archival period. Download work you want to retain and do not upload confidential material unsuitable for processing.

Rights and synthetic media

Upload only material you have permission to process, including any necessary likeness consent. The site does not grant a blanket commercial license or guarantee monetization or ad approval. Check the model license, source assets and destination requirements for your project. Label synthetic or altered footage where appropriate.

Output and product boundaries

Public generation creates silent standard-resolution MP4 clips. There is no public lip-sync, voice-over, precise camera-path editor, cross-task identity lock or upscale switch. The fantasy homepage film was separately upscaled and includes sound. Current delivery adds no brand watermark, though the model can generate lettering or other marks.

A prompt to start with

An explorer stands beneath colossal interlocking orbital towers, luminous bridges disappearing into clouds, slow forward tracking shot, cold rim light, one continuous shot.

Use this as a starting point, then adjust one action or camera instruction at a time. It is a prompt example, not a guarantee of a specific result.

Try this workflow →