100% Private
No Signup
Free Forever
One of 64 free AI tools by Mahmoud Zalt.
Free Depth Map Generator
Turn any photo into a grayscale depth map, fully private|4.8 (469)
A free, private depth map generator that turns any photo into a grayscale depth map without uploading anything. It is built on Hugging Face Transformers.js, the open-source library that runs state-of-the-art machine learning directly in your browser, and it loads Depth Anything, a leading monocular depth-estimation model that infers how far each pixel is from the camera from a single image. Drop in a JPG, PNG, or WebP and the tool runs the model locally, then renders a depth map where bright pixels are near the camera and dark pixels are far away. Because the model runs on your own device via ONNX Runtime and WebAssembly, your image is never uploaded, logged, or stored, which makes it safe for unpublished photos and confidential work. The model downloads once on first use (about 50MB) and is then cached by your browser, so later runs start instantly. You can view the original and the depth map side by side and download the depth map as a PNG for use in 3D, parallax, portrait and background effects, AI art with ControlNet, and photography workflows.
Free and provided as is, without warranty. Use at your own risk. Terms
Generate a depth map from any photo without uploading anything
Most online depth map generators ask you to upload your photo to a server, run it through a hosted model, and send an image back, which means your picture leaves your control and often counts against a paid quota. This free depth map generator works the other way around: the depth-estimation model runs inside your browser tab. Once the model is downloaded, your image is processed on your own device, so the file never touches a server and there is no per-image limit.
That makes it a strong fit for unpublished photos, client work, product imagery, and any picture you would rather not hand to a third party. You still get a standard grayscale depth map PNG you can drop straight into a 3D tool, a compositing timeline, or a ControlNet depth workflow, and you can confirm nothing is uploaded by watching the Network tab in your browser DevTools while it works.
Powered by Depth Anything, running in your browser via Transformers.js
This tool is built on Transformers.js, the open-source library from Hugging Face that runs Transformer models directly in the browser using ONNX Runtime compiled to WebAssembly, with no backend. It loads Depth Anything, a state-of-the-art monocular depth-estimation model, which estimates how far each pixel is from the camera using only a single ordinary photo, with no depth camera, stereo pair, or LiDAR required.
The tool renders the result as a grayscale depth map where bright pixels are near the camera and dark pixels are far away, and shows it beside your original image so you can compare them at a glance. The model files download once from the Hugging Face Hub, are cached by your browser, and are reused on later runs with no repeat download. Both Transformers.js and the Depth Anything model are released under the permissive Apache 2.0 license.
What a depth map is, and how to use one
A depth map encodes distance as brightness: instead of showing color, each pixel shows how near or far that point in the scene is. That single idea unlocks a lot of creative and technical work. In AI art, a depth map is the conditioning image for ControlNet depth, letting you keep the composition and perspective of a source photo while a generator like Stable Diffusion invents new content. In motion and web design, a depth map drives parallax and 2.5D effects, giving a flat image a convincing sense of three dimensions as the viewer scrolls or the camera moves.
Photographers use depth maps to fake portrait-mode background blur on cameras that never captured real depth, and 3D artists use them as displacement or height maps to add relief to meshes and terrain. Compositors turn them into foreground and background masks for selective color grading and edits. Because the depth here is relative rather than metric, it is ideal for these visual uses where the ordering of near and far matters more than an exact distance in meters. Download the PNG, and if a downstream tool expects the opposite convention you can invert it in any image editor before use.
Who reaches for a depth map, and why
AI artists working with ControlNet use depth maps constantly, since a depth-conditioned generation keeps a subject's pose and the scene's composition stable while completely changing style, lighting, or content, which is otherwise hard to control with a text prompt alone. Indie game developers and technical artists use a photo's depth map as a quick starting point for a parallax background layer or a rough height field, without modeling geometry by hand.
Motion designers use depth maps to add a 2.5D camera-push effect to a still photo for a video intro or a social post, and photographers without a portrait-mode phone use one to simulate background blur on an older photo. It also shows up in classrooms teaching computer vision, since it is a rare example of an AI capability, inferring 3D structure from a single 2D image, that is genuinely difficult for humans to do by eye with precision but trivial for a trained model.
How this compares to Photoshop and dedicated depth tools
Adobe Photoshop's Neural Filters include a depth-aware blur that estimates depth internally to fake a shallow-focus look, but it requires a Creative Cloud subscription (around 20 to 55 dollars a month depending on the plan) and does not let you export the raw depth map for use elsewhere. Dedicated depth-capture apps that use a phone's LiDAR sensor produce genuinely more accurate, metric depth, but only work on newer iPhone Pro models and only for photos taken with that specific app.
This tool sits in between: it works on any existing photo from any camera, requires no subscription or specific hardware, and gives you the raw grayscale PNG to use however you want, in ControlNet, in a compositor, or in a game engine. The tradeoff for that flexibility is that the depth is inferred rather than measured, so it is relative and approximate rather than the metric precision a LiDAR capture provides.
The research behind Depth Anything
Depth Anything builds on a lineage of monocular depth research that includes MiDaS, an earlier model that established many of the techniques for training a single network to estimate depth across wildly different photo types. Depth Anything's key contribution was scale: its authors trained on roughly 1.5 million labeled images combined with over 62 million unlabeled images automatically pseudo-labeled by the model itself, using a DINOv2 vision transformer backbone, which is why it generalizes well to photos very different from typical depth-estimation training data.
That scale of training is exactly why a relatively small, quantized version of the model can run inside a browser tab and still produce coherent depth maps across portraits, landscapes, interiors, and product shots without any scene-specific tuning. It is also released under the permissive Apache 2.0 license, which is what makes it legal and practical to redistribute a browser-ready, ONNX-exported version at all.
How It Works
Upload a photo (JPG, PNG, or WebP): nothing is sent to a server, the image stays on your device.
The tool loads the Depth Anything model once on first use, then runs monocular depth estimation locally in your browser.
View the original and the grayscale depth map side by side, where near is bright and far is dark, then download the depth map as a PNG.
Need expert help with AI?
Looking for a specialist to help integrate, optimize, or consult on AI systems? Book a one-on-one technical consultation with an experienced AI consultant to get tailored advice.
Key Features
Privacy & Trust
Use Cases
Limitations
- The first run downloads the Depth Anything model (about 50MB), which is cached by your browser afterward
- Depth is estimated from a single image, so it is relative and approximate, not a metric distance in meters like a LiDAR or stereo camera
- Very large images (over about 1536 pixels on the long edge) take longer and use more memory, since processing runs on your CPU via WebAssembly
- Fine details, thin objects, reflections, transparent surfaces, and heavy motion blur can produce soft or inaccurate depth edges
- Results vary with scene type: clear photos with obvious foreground and background separation give the cleanest depth maps
Frequently Asked Questions
What is a depth map and what is it used for?
A depth map is a grayscale image where each pixel encodes how far that part of the scene is from the camera, so brightness stands in for distance. In this tool, near objects are rendered bright and far objects dark. Depth maps are widely used to create 3D and parallax effects, to drive ControlNet depth conditioning in AI image generators like Stable Diffusion, to simulate portrait-mode background blur, and to build displacement maps and masks for compositing. They give artists and developers a fast way to add a sense of geometry and distance to an otherwise flat photo.
Is this depth map generator really free?
Yes, it is completely free with no signup, no account, and no usage limits. Because the Depth Anything model runs on your own device through Transformers.js instead of a paid cloud API, there are no per-image costs to pass on. You can generate as many depth maps as you want, as often as you want, without a credit card, an API key, or a rate limit.
Is my image uploaded to a server?
No. The entire process, loading the model and running depth estimation, happens locally in your browser using Transformers.js and WebAssembly. After the model is downloaded once, your image is processed on your own device with zero network requests carrying it. Nothing is uploaded, logged, or stored, which makes this safe for unpublished photos and confidential work. You can confirm this by opening the Network tab in your browser DevTools while you generate a depth map.
Which AI model powers this tool?
It uses Depth Anything, a state-of-the-art monocular depth-estimation model, specifically the small variant exported to ONNX for the web. Monocular means it estimates depth from a single ordinary photo, with no special depth camera, stereo pair, or LiDAR required. The model runs through Hugging Face Transformers.js, which executes it in your browser via ONNX Runtime compiled to WebAssembly. Both Transformers.js and the model are released under the permissive Apache 2.0 license.
Q&A SESSION
Got a quick technical question?
Skip the back-and-forth. Get a direct answer from an experienced engineer.