
AI and photography
What each tool does, where it fails, and what it costs.
This is an index of the AI tools a working photographer actually meets, grouped by the job they do. I shoot fashion, portraits and hotels in London, so every note here comes from the work, not from a press release.
Nothing on this website is generated. Every photograph is mine.
| The job | Tools that claim it | Page |
|---|---|---|
| Retouch a whole session | Imagen, Aftershoot, Evoto, Retouch4Me, PortraitPro, ON1, Radiant, Luminar Neo | Retouching |
| Pick the keepers from 1,000 frames | Aftershoot Select, Narrative Select, FilterPixel, Optyx, Excire | Culling |
| Clean noise from a dark frame | Lightroom Denoise, DxO DeepPRIME, Topaz Photo AI | Noise |
| Make a small file printable | Topaz Gigapixel, Let's Enhance, Upscayl, Magnific | Upscaling |
| Cut a subject off its background | remove.bg, Photoroom, Clipdrop, Pixelcut | Background |
| Generate a portrait without a camera | HeadshotPro, Aragon, BetterPic, Remini | Headshot generators |
| Make images from a prompt | Firefly, Flux, Midjourney, Ideogram, Krea, Leonardo | Image generators |
| Turn stills into short video | Runway, Kling, Luma, Topaz Video, Higgsfield | Video |
| Run models on your own machine | ComfyUI, InvokeAI, Upscayl, GFPGAN, CodeFormer | Open source |
Prices and features on this page were read from each vendor's own website on 25 September 2026. Vendors change plans often. Check the current page before you buy.
Each page lists the tools for one job. For every tool you get what it does well, where it fails, and the price the vendor published.
A verdict line marks my own opinion. It is separated from the facts on purpose.
Tool directories change faster than any single page can. For a broader catalogue, the independent directory ToolGlance lists 289 AI tools across 21 categories, including image generation and design.

| Page | Link | What it covers |
|---|---|---|
| AI retouching software | Open | Eight batch tools, what each does to skin |
| AI photo culling | Open | Six ways to pick keepers from a thousand frames |
| AI background removal | Open | Which tool cuts a clean edge, and where every one fails |
| AI photo upscaling | Open | What upscaling adds, and what it invents |
| AI noise reduction | Open | The safest AI job, and how to still get it wrong |
| AI headshot generators | Open | What they cost and the five places they give themselves away |
| AI photo editors | Open | Which part of the work each editor takes over |
| AI image generators | Open | Useful before a shoot, not after it |
| AI video tools | Open | What holds up in client work |
| Open source tools | Open | Free, offline, private, and what you give up |
| AI against manual retouching | Open | Where each one wins |
| AI headshots against studio | Open | The comparison a client actually makes |
| What AI retouching breaks | Open | Ten failures and the fix for each |
| What the stack costs | Open | Ten published prices, three realistic stacks |
| Client files and cloud tools | Open | Five questions before you upload |
Start with the job you repeat most often. If that is culling, fix culling first.
Prefer a tool that runs on your machine when the files belong to a client.
Prefer a tool that learns your edits over one that applies its own taste.
Try the free tier on your three most difficult frames, not on the sample images.
Count the yearly cost before the monthly one, because that is the number you actually pay.
It changed the boring middle of the work: selecting, cleaning, denoising, cutting out.
It changed turnaround. A gallery that took a week can take an evening.
It did not change light, composition, direction or trust.
It did not change who is responsible when a face looks wrong in print.
Most of the value is in the first group, and most of the risk is in pretending it reaches the second.
Faces that will be recognised. Shape and skin stay manual.
Fabric that is the subject of the photograph.
Anything presented as a record: a press image, a product as delivered, a room as it is.
Generated people, at any stage, on a client gallery.
Culling means choosing which frames are worth editing.
Retouching means working on the image itself: skin, dust, shape, tone.
Upscaling means producing a larger file than the camera recorded.
Denoising means removing sensor grain, usually from a RAW file.
Generative means the pixels were invented rather than recorded.
This page is written by Mariya Harley, a fashion, portrait and hotel photographer working from a studio in west London.
The tools described here are the ones that reach a working photographer's desk, and the failures described are the ones that reach a client's screen.
Prices and features are read from vendor pages rather than repeated from other articles. The date of that check is printed above each table.
Nothing here is sponsored, and there are no affiliate links. A tool appears because it does the job, or because people ask about it.
What is deliberately absent: benchmark scores copied from press releases, rankings that change with the season, and any claim that software can replace direction, light or responsibility for the result.
If a price here is out of date, the vendor page wins. Tell me and I will correct it.
Two kinds of people arrive here. Photographers deciding what to buy, and clients deciding whether software can replace a shoot.
Both questions have honest answers, and they are different answers.
For a photographer the question is which hour of the week a tool gives back.
For a client the question is whether the picture has to be true, or only has to fill a space.
Every page below answers one of those two, and says which tools are involved.
Most people who read these pages are deciding one thing: whether software is enough, or whether they need a photographer. Both answers are fine. The pages say which is which.
If the answer is a real shoot, the services and prices page has the numbers, and fashion photography in London explains how I work.
Next step
Send the brief and I will tell you which one solves it.
Six named clients on portrait, fashion, brand, hotel and event work: read the reviews.