Develop
Edit a frame, and it joins the gallery as a picture of its own
Develop edits the photographs already in a gallery, from white balance, tone and crop to masks over part of a frame and dust on the sensor, and puts each edit on the client’s link as a version of its own. There are no layers, and noise is one slider rather than Lightroom’s own, so skin work and heavy noise reduction still go through Lightroom or Photoshop.
Develop is on every plan, the free one included.

What the client opens
Showing a version puts another photograph on the client’s page
Tick Show in the gallery and Latentshoot makes the picture and puts it on the link the client has had since the moodboard. It sits in the grid next to the frame it was made from, and the client can select it and download it without touching the other one.
The client reads DSC_4411 — Black and white, which is the filename with your own name for the version after it, and the file that lands on their disk is DSC_4411-black-and-white.jpg. Nothing on that page names this screen. A black and white and a 4:5 are two photographs as far as the client is concerned, and a select or a note lands on the one they were looking at.
The frame as you delivered it stays in the gallery until you take it out. Show in the gallery is on that one too, so a set of alternates can go out with or without the picture they came from.
What a client can take away is still the gallery’s own download setting, as far as the plan allows. On the free plan a client downloads photographs at web size, and Pro and Studio allow the full-size file as well. What else a gallery holds.
Each version you show becomes a round of that photograph on the client’s page, which they can compare with the last one and approve. A note where they marked an area opens here with their box drawn on the frame. Press Draft a fix and a proposed edit opens here too, over the version they saw, until you keep it or skip it. It sends that picture and the notes to a model and uses 2 AI credits a photograph. How proofing works.
A set at a time
One white balance, pasted onto the frames it belongs on
The strip along the bottom holds the whole gallery, set by set, or one set picked above it. Finish one frame, copy its settings, select the frames the light held on, and paste. The panel asks first, and says which groups of settings it will write and how much each one moves, so the crop and the delivery size can be left out, since those belong to one frame rather than to its colour.
A paste goes onto the version each frame opens on. A frame still on its original gets a new version for the paste, and the original itself is never written. Working through a set in order, Paste settings from previous takes the settings of the frame before the ones selected, so there is nothing to copy first. Each thumbnail in the strip carries a mark: the frame that is open, a frame whose picture was exported before its settings last changed, and a frame nothing has touched yet.
Export this gallery then works through every version with no picture yet, and every one whose settings have moved since the last, one at a time. It says how many it will make and roughly what it will send before it starts. It is a desktop control, because a pass over a gallery can run to gigabytes; on a phone you export one picture at a time.

Where the work happens
Each picture is made on the computer in front of you
The editor runs in the browser, and each picture is made on this device rather than on a server, so keep Develop open while it works. Whatever has finished is kept if you leave, and the rest can be exported again.
While you work, the photograph is developed at the size of your screen, which is what keeps a slider moving under your hand. The panel says as much beside the export button: what is on screen is close, and the exported picture is made again at full size from the same settings. Pressing the button asks first: it says the picture goes to the gallery rather than to your computer, and holds the longest edge, the JPEG quality and the output sharpening, so you can change them before anything is made.
Each picture goes into the same storage as the rest of the gallery, which is 2 GB on Free, 200 GB on Pro and 1 TB on Studio. The whole table is on the pricing page.

RAW files
Edit from the camera’s own file
On Pro and Studio, a photograph in a gallery can keep its RAW file beside the JPEG. Develop opens the RAW file instead, so a version starts from what the camera recorded rather than from a JPEG made from it.
The client still receives the picture you export, and the RAW file stays with your team. A RAW file is larger than a JPEG, so it takes longer to open, and on a phone a full-size export can take a few minutes.
DNG files open too. Develop reads the colour a Leica, a Pentax or a Google Pixel writes into its own DNG, and the colour saved in Apple ProRAW and DxO PureRAW files. A ProRAW file opens on the Camera look profile, which uses Apple’s own tone curve and local tone mapping. Current versions of Lightroom’s Denoise and Enhance save their DNGs with JPEG XL compression, which Develop cannot open yet, and a camera’s small sRAW or mRAW file does not open either.
A phone’s HEIC file can be kept the same way. Develop edits the JPEG made from it, as it would any other picture, and the HEIC stays with the photograph.
How RAW files get into a gallery, from Lightroom Classic or by dropping them on a set, with their JPEGs or on their own.
Part of a frame
Masks you draw or describe, and changes you ask for in words
A mask gives one area of the photograph settings of its own: a gradient across a sky, an oval around a face, or an area you paint with a brush. Inside it are exposure, contrast, highlights, shadows, warmth, tint and saturation, and it stays in place when you crop. Masks are on every plan.
A selection is a mask you describe instead of drawing. Type “the sky” or “the people” and it is outlined for you; the sliders inside it are still yours to set, so nothing in the picture is generated. Each description you send counts as one selection, and one that finds nothing is not counted. Pro includes 300 a month and Studio 1,500.
A change is for something that should not be there. Paint over it, say what should change, such as “Remove this wire” or “Take out the sign”, and the painted area is redrawn with nothing outside it touched. It sits over the photograph like a patch: switch it off to compare, or delete it and the frame is exactly as it was. A change uses 10 AI credits; Pro includes 250 credits a month and Studio 1,000, and one that fails is not counted.
Selections, changes and drafts send a smaller copy of the photograph, or of the area around your paint, to an outside service, which does not keep it. Nothing is sent until the team owner has agreed under Account, in AI permissions, and the gallery allows it in its own settings.
Heal and clone
Paint a wire out or click a speck away, and let Develop find the dust
Click a speck and Develop covers it with a clean patch from nearby. For a longer mark, such as a telegraph wire or a stray hair, paint along it, and Develop finds a clean stretch beside the whole stroke to copy from. Drag the dashed outline to change where either one copies from, and Shift-click to carry a repair on in a straight line, which takes a wire across the frame in two clicks. Heal blends the patch into the light around it, and Clone copies it as it is. Size, softness and opacity are set per repair, and a photograph can hold up to 250.
Find dust looks for soft dark specks in plain areas, such as sky or a studio sweep, and lists each one with a close-up. Nothing comes off until you press Remove. Compare with the other frames in the set, and a speck in the same place on three or more of them is marked as dust on the sensor, which one press removes. Paste the repairs onto the rest of the set from the same camera.
Heal and clone runs on your computer, and so does the dust finder. Neither sends anything or uses AI credits. The finder works on a preview, so check the sky at 100% before a large print.

Round a frame
Extend the frame to fit a different shape
Choose a shape (Square, 4:5, 5:4, 2:3, 3:2, 9:16, 16:9) or set each edge yourself. Drag the photograph to where it should sit and, if you like, say what should be there, such as a clear sky or a plain studio backdrop.
Every pixel of the photograph is kept as it was, and only the room around it is generated. The new area is made at a lower resolution than your photograph and enlarged to fit, so it suits a post or a banner better than a print. Keep several extended frames and use one at a time; As shot gives the photograph back exactly.
Extending uses 25 AI credits and, like a change, needs the team owner’s agreement. What each AI tool sends.

Your usual settings
A starting point worked out from your own edits
Open Your usual settings at the top of the Light and colour panel and Develop offers the white balance, tone, presence, curve, colour, grading and effects you tend to use, taken as the middle value across your own versions. Learn from all your shoots, from one client’s, or from photographs with a tag.
Tick the groups you want, compare on the photograph, and then use them. Exposure and white balance are learned as a figure you set or as a step away from Auto, whichever your edits show. Black and white versions learn only from black and white versions.
It needs twelve edited photographs across three shoots before it offers anything. It is worked out on your computer, sends nothing and uses no credits.

Presets
Your own presets, and the ones you already have
Save the settings you have arrived at as a preset and it is held for the team, so what you worked out on one job is there on the next. Applying one shows you which groups of settings it will write before it writes them, so you can take its tone curve and leave its white balance. You rename, file and delete them under Account, in Presets.
Bring your own .xmp presets in. Latentshoot maps what it has a setting for and tells you what it could not carry, so check each one on a frame before you rely on it.
Match a reference starts from a photograph instead: pick one with the colour and tone you are after, and Latentshoot suggests settings to match it, which you can keep in part and save as a preset. It uses 6 AI credits.
Where it stops
What still goes through Lightroom
There are no layers. Heal and clone takes dust and wires off by copying the photograph’s own pixels, and a change can take out something the frame has nothing to cover with, but a change is a model redrawing the area you painted, so look at it at full size before it reaches a client. Skin work, and anything a client will study closely, still belongs with whoever retouches for you now.
Noise is one slider, over the luminance. A frame needing the kind of noise reduction Lightroom’s own does still goes there.
Your catalogue stays where it is. Latentshoot holds the photographs you have delivered rather than your library, so there is no importing, no tethered capture and no printing. Publishing a collection from Lightroom Classic is how most of them arrive.
The client asks, and you make the change. They can mark an area with a note and approve a round, but nothing they do edits a photograph. A draft from their notes is a proposal you keep or skip, and they see only what you have shown in the gallery. How proofing works.
Questions
About Develop
Does this replace Lightroom?
Which files can I edit?
What happens to the file I uploaded?
Can the client see a version before I am finished with it?
Does editing use my storage?
Does taking dust off use AI credits?
Where do my usual settings come from?
What counts as a selection?
Is anything about a photograph generated?
Can a client’s notes start an edit?
Can I edit on a phone?
Plan the shoot, then edit what came off it
The free plan carries one shoot end to end, from the moodboard to the photographs the client downloads.
No card.