DxO PhotoLab?

NOT REALLYA prototype of the interface is possible, but the durable product depends on difficult systems beyond the UI. Build a safer adjacent workflow instead.

BUILD FIRST

A non-sensitive or read-only version of the core experience
A manual review step wherever correctness or trust matters
A narrow companion tool that removes a real piece of work

LEAVE FOR LATER

Regulated decisions, payments, or irreversible actions
Claims of correctness, compliance, or uptime that the MVP cannot support
Sensitive integrations until security, support, and auditability are ready

RESEARCH NOTES

What this tool actually does.

Checked 2026-09-06

DxO PhotoLab is a desktop, non-destructive RAW photo editor from French company DxO (Windows/macOS; current major version 9). It centers on DxO Modules: per camera/lens optical corrections (distortion, vignetting, lateral/longitudinal chromatic aberration, sharpness) measured in DxO labs and applied automatically at import. Machine-learning DeepPRIME/DeepPRIME XD denoising removes noise while recovering detail and color, with a conventional 'Standard' mode for speed. Featured tools include ClearView Plus (haze removal), Smart Lighting (shadow/highlight balancing), U Point Control Points/Control Lines local adjustments, Brush/Gradient/Eraser/Auto masks, Hue Masks, AI Masks for sky/face/hair/subject, ColorWheel, tone curves, film-style renderings, and DCP profile import. It has a full History palette, soft-proofing in the wide-gamut DxO Wide Gamut color space, parallel exports (JPEG/TIFF/D

SOURCES

RESEARCHED BUILD PROMPT

Give this to your coding agent.

Build a web-based RAW photo editing workbench inspired by the workflow of DxO PhotoLab 9 (desktop raw editor). Do not copy DxO's brand, name, logos, proprietary assets, or UI; create an original interface and name.

Core workflow to replicate: 1) Import RAW or JPEG; for RAW, run a demosaicing step before edits. 2) Auto-apply optics-profile-style corrections (distortion, vignetting, chromatic aberration) from a small built-in profile table users can extend via JSON. 3) Global adjustments: exposure, contrast, highlights/shadows, white balance with an eyedropper picker, tone curve, saturation/vibrancy, and a black-and-white channel mixer. 4) Local adjustments using brush, gradient, and eraser masks: paint a mask effect, edit only that region, plus click-point region growing (U Point-style Control Points) and an optional client-side segmentation fallback for sky/subject. 5) A machine-learning denoise slider (ONNX/WASM model). 6) Non-destructive edit stack with a history panel, before/after toggle, and export to JPEG/PNG.

Technical requirements: React + TypeScript front end; preview rendering via WebGL or WebGPU; non-destructive edits stored as a serializable JSON edit list kept in sync with the renderer; RAW handling via a WASM demosaic library (e.g., libraw port) supporting at least a few sample formats; exports via OffscreenCanvas; file access via the File System Access API with a normal file-input fallback.

Explicitly out of scope: a lab-calibrated camera/lens database beyond demo profiles, DAM/library management, print-accurate ICC soft proofing, plugin ecosystems, and any claim of production-grade color science fidelity.

Deliver a README covering setup, dependencies, and known limitations (raw format coverage, model accuracy, performance on large files).

Visit the official product page