Nano Banana Face Swap Prompt: 15+ Ready-to-Use Prompts for Realistic Results
A Nano Banana face swap prompt tells the AI which face to take from one image and where to place it on another. The simplest working prompt is: “Insert the face of the person from image one into the body of the person of image two.” You upload two images to Nano Banana in a specific order — face first, body second — and the model handles lighting, skin texture, and bone structure automatically.
This guide covers the exact prompts for professional headshots, cinematic scenes, artistic styles, and travel compositions, plus the upload order and realism tips that make the difference between a convincing face swap and an obvious fake.
How Nano Banana Face Swap Works
Unlike older face swap tools that paste a cut-out face onto a target head, Nano Banana understands 3D geometry. Google DeepMind built the model as part of the Gemini family, and it processes face swaps differently from traditional pixel-matching approaches. Instead of simple overlay, the AI-powered image editing model analyzes the bone structure of both faces, maps the source identity onto the target pose, and then re-renders lighting, skin texture, and shadow angles to match the target environment.
This process maintains 95% character consistency across edits. Eyes, eyebrow shape, jawline, and skin tone from the source face transfer accurately even when the target body is in a completely different lighting setup or camera angle.
The distinction between face swap and head swap matters for choosing the right prompt. A face swap replaces only the facial features — eyes, nose, mouth, cheeks, skin — while preserving the target image’s head shape, hair, and body. A head swap replaces everything from the neck up, including hair and skull shape. Nano Banana Pro handles both, but the prompts differ significantly.
Step-by-Step: How to Face Swap with Nano Banana
The process takes under 60 seconds once you know the correct sequence. Image upload order is the single most important factor — get it wrong and the AI swaps the wrong face.
- Open the Gemini App and make sure you are using Nano Banana Pro (the thinking version) — the regular Nano Banana model produces unreliable face swap results
- Upload your source face photo as the first image (Image 1) — this is the face you want to transfer
- Upload the target body/scene photo as the second image (Image 2) — this is where the face will be placed
- Paste the prompt: “Insert the face of the person from image one into the body of the person of image two”
- Click “Create images” and wait 3-5 seconds for the result
If the initial result needs refinement, add descriptive details to help the model identify each person. For example: “The person in Image 1 is an Asian woman with dark hair. The person in Image 2 is a blonde woman in a red dress. Swap the Asian woman’s face onto the blonde woman’s body while preserving the blonde woman’s hair and outfit.” This level of specificity reduces confusion and improves accuracy.
15+ Ready-to-Use Nano Banana Face Swap Prompts
Every prompt below follows the Multi-Image Reference format: Image 1 is always your face source, Image 2 is always the target scene. Copy any prompt directly into Gemini.
Professional and Corporate Headshots
These prompts produce LinkedIn-ready portraits and team page photos. The key is specifying studio lighting and skin detail.
“Using the face from Image 1, swap it onto the subject in Image 2. Maintain a professional, warm smile. Adjust the lighting to a soft-box studio setup with a shallow depth of field (f/1.8). Ensure the skin grain and pores are realistic.”
“Transfer the facial features from Image 1 to a subject wearing a beige turtleneck against a charcoal background. Focus on hyper-realistic skin texture and eye reflections. Nano Banana Pro 4K resolution.”
“Replace the subject’s face in Image 2 with the face from Image 1. The character is sitting in a modern glass-walled office at sunset. Golden hour light hitting the left side of the face.”
“Swap the face from Image 1 onto the body of a speaker on a high-tech stage. Keep the jawline and eye shape 100% consistent with Image 1. Lighting should be cinematic stage-blue with rim light on the shoulders.”
Cinematic and Movie-Style Prompts
For dramatic compositions with environment-specific lighting that interacts with the swapped face.
“Insert the face from Image 1 into a rainy, neon-lit street scene. The subject is wearing a high-collar leather jacket. The face should reflect the pink and cyan neon lights from the environment.”
“Swap the identity from Image 1 onto a warrior in ornate silver armor. Maintain the exact eye color and eyebrow shape. The setting is a misty forest at dawn. Epic, cinematic color grading.”
“Apply the face from Image 1 to a hard-boiled detective in a rainy alleyway. Style: Film Noir, high contrast, dramatic shadows (Chiaroscuro). The face must look like a high-grain vintage photograph.”
“Swap the face from Image 1 into a detailed astronaut suit. The helmet visor should show a faint reflection of a distant galaxy. Preserve the facial bone structure perfectly.”
Artistic and Stylized Prompts
Style transfer prompts that preserve facial identity while applying a completely different artistic medium.
“Reinterpret the face from Image 1 as the subject of a 16th-century oil painting. Maintain the likeness but adapt the skin tones to a sfumato technique. Heavy impasto textures on the clothing.”
“Convert the facial identity of Image 1 into a clean, minimalist vector art style. Bold lines, flat colors, but keep the recognizable features of the person.”
“Using Image 1 as the reference, create a hyper-realistic white marble bust. The facial features must be carved with precision, showing subtle cracks and stone texture.”
Travel and Lifestyle Prompts
These prompts place a face into location-specific environments with realistic environmental effects on the skin.
“Place the face from Image 1 onto a person enjoying an espresso on a balcony overlooking Positano. Bright, saturated Mediterranean colors. Natural sunlight.”
“Swap the face from Image 1 onto a subject walking through Shibuya Crossing. High-speed shutter look, vibrant street lights, and realistic motion in the background.”
“Identify the face in Image 1 and swap it onto a hiker standing on a snowy peak. The face should show subtle redness from the cold and realistic wind-blown hair.”
| Category | Best For | Key Prompt Elements |
|---|---|---|
| Professional | LinkedIn, team pages, headshots | Soft-box lighting, f/1.8 depth of field, skin pores |
| Cinematic | Movie scenes, fantasy, sci-fi | Rim lighting, color grading, environment reflections |
| Artistic | Social media avatars, prints | Style name (sfumato, vector, marble), medium texture |
| Travel | Lifestyle content, vacation edits | Location-specific light, weather effects on skin |
Tips for Hyper-Realistic Results
Getting a face swap that looks convincing rather than obviously AI-generated comes down to three factors: source image quality, prompt specificity, and model selection.
Start with a 2K+ resolution reference photo. The face source (Image 1) determines output quality more than anything else. Use a photo with neutral, even lighting where the face is fully visible — no sunglasses, hats, or extreme angles. Face-forward shots with both eyes visible produce the most consistent results.
Explicitly mention micro-details in the prompt. Add “preserving skin pores, natural imperfections, and original eye color” to any face swap prompt. Without these instructions, Nano Banana tends to smooth skin and brighten eyes, which creates an uncanny valley effect. Including imperfections makes the result look human.
Add identity descriptors for each person. Describe both subjects: “The person in Image 1 has brown eyes, freckles, and a narrow jawline. The person in Image 2 has a wider face and olive skin.” This helps the AI correctly separate which features to keep and which to replace.
Use Nano Banana Pro, not the regular model. The free-tier Nano Banana (Gemini 2.5 Flash Image) handles basic image editing but struggles with complex face swaps. Nano Banana Pro (Gemini 3 Pro Image) delivers dramatically better results thanks to its 3D geometry understanding and higher character consistency.
| Feature | Nano Banana | Nano Banana Pro | Flux Kontext |
|---|---|---|---|
| Face swap quality | Basic | Professional | Good |
| Generation speed | ~0.8s | ~2s | 10-15s |
| Character consistency | ~85% | 95%+ | ~71% |
| Max resolution | 1MP | 4K | 2K |
| Multi-image inputs | 1 | Up to 14 | 1-2 |
| Price | Free (3/day) | $19.99/mo | Varies |
Pricing: Free vs Pro for Face Swaps
Google structures Nano Banana access across tiers that directly affect face swap quality and volume.
Daily Face Swap Limits by Subscription
- Free tier — 3 face swaps per day at 1MP (1024×1024). Visible Gemini sparkle watermark. Good enough for testing prompts but resolution is noticeably lower than Pro
- Pro subscription ($19.99/month) — approximately 100 images per day at up to 4K resolution. This is the recommended tier for regular face swap use
- Ultra subscription ($124.99/month) — approximately 1,000 images per day. Full 4K, no visible watermark. Designed for production and commercial workflows
- API access (~$0.15 per 4K image) — available through Google AI Studio or Vertex AI. Pay-per-generation, no visible watermark, up to 16MP output
Limitations and Ethics of AI Face Swap
Nano Banana face swap has clear boundaries that are important to understand before building a workflow around it.
Video face swapping is not supported. Nano Banana processes still images only. For video face replacement, Google’s Veo model is the intended tool — you can create consistent keyframe images with Nano Banana and then animate them through Veo’s image-to-video pipeline.
Complex angles produce lower quality results. Profile views, looking-down poses, and faces partially hidden by hands or objects all reduce swap accuracy. The model works best with face-forward or slight three-quarter angle photos where both eyes are visible.
SynthID embeds a digital watermark directly into the pixels of AI-generated images, making it imperceptible to the human eye while remaining detectable by automated tools.
Google DeepMind — SynthID
Every face swap generated by Nano Banana carries an invisible SynthID watermark. This watermark survives cropping, resizing, and JPEG compression. Only perform face swaps on images you have rights to use. Creating non-consensual deepfakes of real people may violate local laws, and most platforms include safety filters that block attempts to generate them.