A depth map is a grayscale image where each pixel encodes the estimated distance from the camera. Bright areas are closer, dark areas are farther away. Depth maps are used in 3D compositing, blur effects (bokeh), and visual effects.

What model does this use?

Depth Anything V2 Small by ByteDance, running via @huggingface/transformers (transformers.js) in your browser. It is a monocular depth estimation model trained on a large and diverse dataset.

Is processing done on my device?

Yes, entirely. The model runs locally using WebAssembly. Your photos never leave your browser.

How accurate is monocular depth estimation?

Monocular depth estimation is relative, not metric — it cannot give you real-world distances in metres. It excels at showing the relative depth order of objects (which is nearer, which is farther) and is very useful for artistic effects and VFX compositing.

What can I do with the depth map?

Common uses include: creating bokeh/lens blur effects in Photoshop, driving depth-of-field in 3D software, creating parallax animations, or generating 3D photo effects on social media.

AI Depth Map Generator

Generate a depth map from any photo using Depth Anything V2, running entirely in your browser. Useful for 3D effects, post-processing, and visual effects. 100% free, no upload.

How it works

1Upload any photo by dragging it in or clicking to browse. Works with JPG, PNG, and WebP.
2On first use the Depth Anything V2 Small model (~50 MB) downloads once and caches in your browser.
3Click "Generate Depth Map" — the model estimates depth for each pixel entirely on your device.
4See the depth heatmap displayed alongside the original. Closer objects appear brighter, farther objects appear darker.
5Download the depth map as a PNG to use in photo editors, 3D software, or video effects.