One photo: each tool needs exactly one reference photo. The AI character generator is the exception: it can start from a description alone or from several images.

Optional direction: leave it blank for the tool's default result, or type one short sentence to steer it.

One result by default: to get a set, include a number from 2 to 10 in the direction, such as "three angles: one from the left, one from the right, one from above." The results arrive as one carousel post and each image is charged at the normal per-image price. AI image to video always creates one clip.

Model and credits: image tools default to GPT Image 2 and AI image to video defaults to Veo 3.1 Fast. The normal credit price for the selected model and settings applies.

Aspect ratio: image tools default to 4:5 (832 x 1,040 pixels at 1K). Profile picture is square (1,024 x 1,024) and AI image to video is 9:16. Attaching a photo no longer changes the ratio; the tool default, or the ratio you pick under Settings, is used for every upload.

Your whole photo is kept: when the chosen ratio differs from your upload, image tools extend the surroundings from what the photo shows instead of cropping, stretching, or letterboxing it. For a landscape result, choose 3:2 or 16:9 under Settings before creating. Profile picture composes a new, tighter square, and AI image to video center-crops the photo because it becomes the exact opening frame.

Edit prompt: restores only the direction you typed. If you left it blank, it stays blank. Typeshutter's private tool instructions are never shown.

Signed-out drafts: you can choose a photo and set up a tool before logging in. The draft stays in that browser for up to 24 hours, and nothing is uploaded or charged until you select Create after logging in.

Characters: tools do not work with a saved Character. Choosing a tool removes the selected Character and shows a notice.

Fixes: temporary blemishes and redness, oily shine and flash hot spots, temporary under-eye darkness, stray and flyaway hairs, glare on eyeglass lenses, lint, dust, crumbs, and smudges, and a color cast or one face falling darker than the others.

Keeps: identity, head pose, gaze, apparent age, skin tone and skin brightness, expression, hairstyle and hairline, pores, wrinkles, freckles, moles, scars, stubble, gray hair, the shape of every feature, and the background.

Directions: use retouching words such as shine, breakout, glare, flyaways, lint, or color cast. Chin softening happens only when you ask, and only through light and shadow. Requests to slim, lift, de-age, add makeup, or change eye color are ignored.

Tip: image models can still lift a deep complexion slightly. Starting the direction with "keep my skin tone exactly as it is" is the surest way to hold it.

Result: film color, soft highlight roll-off, organic grain that is visible at normal viewing size, slightly softer micro-contrast, and warm halation only around bright highlights already in the photo. It is a clean lab-scan look, not a subtle preset and not a distressed vintage overlay.

Keeps: the full frame, every person and face, exposure and brightness, white balance, lighting direction, focus, and perspective.

Never adds: light leaks, scratches, dust, date stamps, sprocket holes, borders, fade, heavy vignette, relighting, or retouching.

Directions: describe the film character, such as "warmer and more saturated," "cooler with greener shadows and softer contrast," "more grain," "fine grain only," or "no halation." That works better than naming a brand of film. Typeshutter is not affiliated with any film manufacturer.

Leaves out: the phone in a mirror selfie, an outstretched arm, a selfie stick, mirror edges, wide-angle selfie distortion, a raised glass, and other people at the edge of a group photo.

Keeps: your face, apparent age, skin tone and real skin texture, glasses, hair, every garment and accessory as worn, and the same place and light.

Never adds: a suit, a studio backdrop, a beauty filter, or new lighting.

Directions: "Keep the laugh," "No smile," "Keep the dog in the picture," or "Come in close and keep the mountains behind me." A look off-camera or a side profile happens only when you ask.

Size: 1,024 x 1,024 pixels, which fits LinkedIn, X, Instagram, and Discord profile pictures.

Keeps: the number of people, faces, hands, clothes, glasses, jewelry, tattoos, head coverings, hair, body size, and pets or objects someone is holding. A kitchen stays a kitchen and an office stays that office.

Never adds: a suit, a studio, a beauty filter, a new location, or a lower body the photo did not show. A waist-up photo gives a waist-up or closer portrait.

Directions: "Come in closer: a head-and-shoulders portrait," "Keep the mug of tea and the washing line," or "Everyone from the chest up."

Groups: one photo is enough, but everyone must be in it with every face visible.

Blank direction: a campaign-style ad image suited to the product. Jewelry, watches, eyewear, bags, and clothing get an editorial close-up worn by a model, or the piece alone on a surface such as stone, silk, or plaster. Drinks, food, cosmetics, and other packaged goods appear as the hero on a gradient in the product's own brand color, with studio light and details from the product itself.

White background: type "white background," "Amazon listing," or "packshot" to get a clean pure-white product photo with no props for marketplaces.

Lifestyle scene: describe it in one line, such as "On a dark slate table by a kitchen window, morning light."

Keeps: the same item, the count of items, colors, every word on the label, the fill level, and the closed, folded, or sealed state.

No text is added: the image contains only the lettering on the product. To add a headline, type the exact words in quotation marks. Words like ad, campaign, poster, billboard, banner, or story describe the look and format; they do not put the product on a billboard unless you ask for that.

Ratio: pick the ratio for the format you need under Settings, for example 4:5 for a feed post, 9:16 for a story, or 1:1 for a marketplace listing.

Check the label: small print that is unreadable in your photo stays soft instead of being invented. Review label text before publishing, and tell marketplaces that ask for AI-image disclosure.

Blank direction: the tool picks one clear change, for example another season or time of day for an outdoor scene, a different garment color for a person, or a different surface and setting for a product.

Directions: replace an object; change a garment, color, material, weather, season, time of day, pose, expression, or background; or add or remove a named element. Everything you do not mention stays the same.

Not for camera moves: to see the scene from another angle, use AI camera control.

Directions: write one sentence naming a camera move, such as "Orbit left," "Closer, from a low angle," "Higher, looking down at the table," or "From the doorway, looking back toward the sofa." Anything that is not a camera move, such as a color, a season, or a smile, is left undone.

Blank direction: the tool picks one bold move on its own.

Keeps: the people and their faces, poses, and outfits, the objects, the room, the light, and every word on every label.

What the move reveals: it is continued plainly from what the photo shows, such as the same brick around the corner or the same floor under the camera. The tool does not furnish a wall or doorway the photo did not show.

Limits: a move that has to reveal a lot the photo never showed, such as the back of a house or a face turned fully away, can look bare. The model sometimes lands a shorter orbit than asked; a second run or naming a spot in the scene usually fixes it. A tightly cropped flat object or a screenshot gives the tool little to work from.

On posts: a public post created with a tool shows a Made with row in its details, with the tool's name and a Try it button. Other people see it, signed in or not; you do not see it on your own posts. Posts made in the standard Create composer without a tool do not show it.

Your direction stays private: the row names the tool only. It never shows the prompt or direction you typed.

On tool pages: each tool page can show a Made with this tool section with up to 12 recent public posts from featured community profiles. A post can appear only while it is public, finished, and not removed or under review. Sensitive posts follow each viewer's sensitive-content setting.

Opting out: there is no setting to hide the row. Make the post private or delete it to remove the association.

My landscape photo came back vertical: image tools default to 4:5 and no longer pick the ratio from the upload. Choose 3:2 or 16:9 under Settings and create again.

I wanted a white product background: add "white background" or "Amazon listing" to the direction and run it again.

There is no headline on my product ad: the tool adds no text unless you type the exact words in quotation marks.

The camera control result looks like my upload: name a specific move or a spot in the scene, such as "From the doorway, looking back toward the sofa," and run it again.

A face, skin tone, label, or outfit changed: run it again with a direction that says what to keep. If it keeps happening, contact Support with the media link and the direction you typed.

The Made with row is missing: you are looking at your own post, or the post was not created through a tool.

My post is not on the tool page: the section shows a small selection of recent public posts from featured profiles, so most posts do not appear there.