LOCAL AI · PERSONAL RAW EDITING
DIRSCHL AI Photo
DIRSCHL AI Photo learns a personal look locally and brings it directly into professional RAW workflows.
In active development · macOS first · Paid desktop app planned · No public download yet

Your editing style – learned locally.
A shared DINOv3 image encoder understands visual content. Personal editing profiles and native control adapters remain strictly separate.
Local learning
RAW files, edits, and training data remain on the user’s own computer.
Inside the RAW workflow
Editing, culling, and project training start from the supported photo applications.
Explicit approval
Only explicitly confirmed final images enter profile training.
One visual style, native controls.
Adobe Camera Raw, Capture One, and darktable render controls differently. Their sliders are never equated; translations are calibrated from rendered images.
Adobe Bridge & Camera Raw
XMP-based editing, protected batch processing, optional AI denoise, and automatic alignment.
Capture One Pro
An independent macOS adapter using the documented AppleScript interface and native Capture One controls.
darktable
A free alternative with its own Lua integration, native styles, and local profile training.
Lean in daily use. Controlled in detail.
Culling with clear logic
2 for closed eyes or blur, 3 for similar images, and 4 for the best image in a series. 1 and 5 remain manual.
Existing work stays protected
Images with existing edits or ratings are not silently overwritten.
Adjustable performance
Local training adapts to the computer and can run in the background without cloud compute costs.
Stay offline or share the style.
DAIS Profiles are planned as compact, signed .daisprofile files. The model is shared – never the RAW files, client data, or local training database.
DIRSCHL AI Photo grows from a real photography workflow.
Pricing, licences, and public test access will be announced after quality and security checks are complete.
