# AI 90s Photo Generator > Upload one portrait, choose from six written 1990s scene directions, or use Custom to write a direction or add one style reference photo. Choose framing and 1K, 2K, or 4K resolution to generate one image with GPT Image 2.5. The model aims to keep the person recognizable, but details can vary. Canonical URL: https://imaginescenes.com/ai-90s-photo-generator ## New-account credits - New accounts receive 40 free credits after signup with Google or GitHub. Use them for generation within 30 days of signup. ## How the 90s photo generator works 1. Upload one clear portrait. It is the person reference. 2. Choose one of six scene cards; each sends a written direction, not its example picture. Or open Custom to write your own direction or upload one style reference photo. Custom text and a style photo are mutually exclusive. 3. Choose an aspect ratio and 1K, 2K, or 4K resolution, check the displayed credit cost, and generate one image. 4. Review the result, then download it or find it in My Creations. ## Scene examples Yearbook, Mall booth, House party, Road trip, Roller rink, and Record store remain available as presets. Four additional illustrative source/result pairs show yearbook, roller-rink, mall-booth, and music-zine moods with different fictional subjects. A scene card sends only its written prompt and your portrait. These pictures are examples, not authentic archival photos or your result. ## Capability boundaries The model is asked to preserve your recognizable face while changing the setting, lighting, wardrobe direction, and film-camera treatment. It may still alter facial details, hair, hands, clothing, or the background. A style reference guides the look and is not a second person to copy. ## Frequently asked questions ### What photo should I upload? Upload one recent, clear portrait with a single visible face and even light. Your photo is the person reference for the result; a simple pose gives the model a clearer starting point. ### How do the six scene cards work? Choose a scene such as Yearbook, House party, or Roller rink after uploading your portrait. Each card supplies a written 90s direction; its example picture is for browsing and is not sent as a second image. ### Can I use a written direction instead of a style photo? Yes. Open Custom and write a short 1990s scene direction, such as a 1996 mall photo booth with direct flash and faded consumer film. A scene name, light, setting, and camera feel are useful details. ### How does a style reference photo work? Open Custom and upload one style photo instead of a written direction. The first image remains your portrait; the second guides color, setting, wardrobe direction, lighting, and film-camera treatment. The model is asked not to copy a person from that photo. ### Can I use both a prompt and a style photo? Choose one direction at a time. The editor clears the other direction when you switch modes, and the service rejects a request containing both. ### Will the result look exactly like me? The workflow asks the model to preserve the recognizable person in your portrait. AI generation can still change facial details, hair, hands, clothing, or the background, so review each result before sharing it. ### Can I make several looks with the same portrait? Yes. Keep your portrait and make separate generations with different scenes or Custom directions. Each run produces one image and uses the credit cost shown in the editor before you submit it. ### Does a scene card use the person shown in its picture? No. Scene pictures illustrate the visual direction. A scene card sends its written description with your portrait; the model is asked to use the person in your upload. ### Will the picture match a real 1990s photograph? The model can imitate visual cues such as direct flash, film grain, and faded color. It does not recover an actual historical photograph or guarantee period-perfect clothing, objects, or camera artifacts. ### Which framing and resolution can I choose? Choose an aspect ratio such as 4:5 for a portrait, 1:1 for a square image, or 3:2 for a camera-style frame. Resolution options are 1K, 2K, and 4K; the editor updates the credit cost when you change resolution. ### How many photos does one generation make? Each submission requests one finished image. You can make another version with a new direction or different framing; each run has its own credit cost shown before you generate. ### Where can I find or save my finished photo? The result appears in Output when generation finishes. You can download it there or open My Creations later to revisit it. ### What if my first result misses the look I wanted? Try a more specific Custom direction, change the scene, or use one style reference with clearer lighting and setting. Review the credit cost before generating another image. ## Related photo tools - https://imaginescenes.com/virtual-hairstyle-try-on - https://imaginescenes.com/ai-cartoon-avatar-maker - https://imaginescenes.com/ai-background-changer - https://imaginescenes.com/ai-age-transformation Updated: 2026-09-26