A close-up of a stylus retouching an AI-generated car image on a pen display

Chapter 01

Cleanup

The last ten percent, by hand.

Beauty, cleanup and compositing on stills and frames. Skin, product, type and detail corrected until the image can actually be signed off.

A plain grey placeholder box on a wooden table in a warm interiorA cream ceramic vase on a wooden table in a warm interiorA dark sculptural vase on a wooden table in a warm interior

Chapter 02

AI Retouch

Cloud or local — the material decides.

Which model, how to steer it, when to take over by hand. Recasting, restyling, relighting, versioning and localisation. Anything under NDA runs through ComfyUI workflows locally on my own machine.

A woman in an orange rain jacket on a pier, the original 4:5 frame outlined in red and the extended sky and sea hatchedThe finished 16:9 version: the woman on the pier with open sea on both sides

Chapter 03

Formats

One key visual, every format.

16:9, 9:16, 4:5 and out-of-home from a single approved image. Extended and recomposed for each ratio — not cropped, not stretched — so every placement looks shot for it.

A film-studio set with a green screen behind a living-room sceneThe same living-room scene at night, the set extended into a room with a waterfront city skyline behind the windows

Chapter 04

Set extensions

Build what wasn't on set.

Rooms, architecture and environments extended beyond what was built on set. 3D and AI matte painting, matched to the plate and tracked to the camera.

Booking

How a job runs.

01

Cleanup

Send the frame that's nearly right and tell me what it should be. Skin, product, cleanup — I fix it by hand, layered and non-destructive.

02

AI Retouch

Send the approved shot and what has to change — the actor, the outfit, the light, the market. I run the models and hand it back retouched.

03

Formats

Send the approved key visual and the list of placements. I extend and recompose it for each ratio, keeping space where the layout needs type.

04

Set extensions

Send the frame and tell me how far the set should reach. I extend it in 2D, 3D or a mix, built on your camera data, so the perspective holds.

Send me your request and deadline, get a figure within 24 hours.