Live AI rendering
Send one moment through time.
Upload a still portrait, choose a past or future direction, and render a cinematic five-second Echo. TimeBloom uses fal.ai's Kling image-to-video model; it does not fabricate a 15-second clip.
JPG, PNG, or WebP only, up to 10 MB. The current model accepts a still image, not a source video.
Each guest browser session can render one real Echo. Your final video is served by the rendering provider and can be downloaded or shared.
Guest Echo studio
One real five-second cinematic render from your uploaded image.
Add a still image
Starting a render uploads your selected image to TimeBloom storage so it can be sent to fal.ai. Provider output URLs are shareable; do not upload a private image you would not want processed.