How to use this prompt
Copy the prompt and replace any placeholders with your reference images or desired elements. Paste it into your AI image generator (MidJourney, DALL·E, or Stable Diffusion) and render in 9:16 aspect ratio to preserve the iPhone selfie composition. Emphasize features like slight blur, grain, flash, and soft focus to achieve the authentic lo-fi look. Add optional accessories like eyeliner, jewelry, or stickers to match your style.
Prompt
A 2010s iPhone selfie of Kirsten Stewart with a lo-fi background, with me ( as same facial structure from the refernce image ) on Billie Eilish's shoulder, my expression blank, with no clear subject or particular composition—just an unintentional candid shot. Slightly blurry. Flash photography. Grainy texture. Edited and soft focus. Capturing an imperfect moment. Uses (eyeliner, jewelry, face stickers, etc.)
Why this prompt works
9.16 aspect ratioThis prompt leverages AI’s ability to mimic spontaneous, imperfect photography from the 2010s. By including slight blur, grainy textures, soft focus, and unintentional composition, it produces images that feel authentic and nostalgic. Adding a real-world reference for your facial structure ensures personalized and recognizable results. The 9:16 aspect ratio mirrors modern phone selfies, making the output perfect for social media or digital art projects.
Example Output
Here's an example result generated with this exact prompt:
Prompt Variations
- Change “Kirsten Stewart” to another celebrity or character → get a personalized candid selfie with anyone
- Change “slightly blurry” to “more motion blur” → get a more chaotic, dynamic candid effect.
- Change “flash photography” to “natural daylight” → get a softer, casual selfie vibe.
- Change “lo-fi background” to “urban street background” → get a trendy, city-style candid shot.
FAQ
Can I use my own photo in this prompt?
Yes — upload your reference image to ensure the AI matches your facial structure accurately in the selfie.
Will the AI always include Billie Eilish and Kirsten Stewart?
It depends on the AI model. If one element is missing, emphasize both characters in the prompt or try separate passes for better control.