Depth Studio
Turn a photo 3D — AI depth, bokeh, relight & parallax. On-device.
Runs entirely in your browser — nothing you enter is uploaded, logged, or tracked.
A real depth-estimation neural network (Depth Anything V2) runs 100% on your device and builds a 3D depth map of any photo — no upload, no servers. With that depth, Depth Studio gives you true depth-aware tools: realistic portrait bokeh where the blur grows with distance (not a flat blur), a relight mode that lights the scene from a depth-derived surface, and a 3D-parallax “moving photo” you can export as an MP4 or GIF. The AI model downloads once (~49 MB) and then works offline forever; your photo never leaves your device, and every export is rebuilt from raw pixels so GPS and camera metadata are left behind.
Export
Rebuilt on your device — no GPS or camera metadata, no watermark.
Frequently asked questions
Is my photo uploaded anywhere?
No. The depth AI and all the effects run inside your browser on your own device — there are no servers, no upload, no analytics. After the one-time model download you can disconnect from the internet and it still works.
What is a depth map and how does an AI make one in my browser?
A depth map estimates how far each pixel is from the camera. Depth Studio runs a compact depth-estimation neural network (Depth Anything V2 Small) compiled to run on your device via WebAssembly — the first time you use it the ~49 MB model downloads once, caches, and from then on every depth map is computed locally and offline.
How is the bokeh different from a normal background blur?
A normal blur is flat — everything behind the subject gets the same blur. Depth-aware bokeh uses the depth map so the blur grows with distance, just like a real camera lens, and you can set the focal plane (what’s in focus) and the aperture (how strong the blur gets). The result looks like a portrait shot on a fast lens.
What is the 3D-parallax / “moving photo”?
It animates a subtle virtual camera move and shifts each pixel by its depth — closer things move more than far things — so a still photo turns into a short 3D shot. Export it as an MP4 (where your browser supports it) or a GIF that plays anywhere.
How well does it work, and what’s best?
It’s strong on photos with clear depth — portraits, landscapes, objects with foreground and background. Very flat scenes or busy edges (hair, fences, glass) are harder for any depth model, and the parallax can stretch slightly where the camera “sees behind” a near object. It’s an on-device model, trading a little of a giant cloud model’s peak quality for total privacy and offline use.
Does it remove EXIF / GPS metadata?
Always. Every export is re-encoded from raw pixels, so the original’s GPS location, camera model and EXIF tags can’t carry over.