Reimagine
Reimagine is an on-device Gemini Nano AI image generation feature that transforms existing photos into new artistic interpretations using style transfer and generative models.
Why It Exists
Users want to creatively reimagine their photos without complex editing software or cloud-based AI services that compromise privacy.
How It Works
Reimagine uses Google's Gemini Nano generative AI models to reinterpret existing photos with different artistic styles, lighting, color palettes, and compositions. The system can transform a daytime photo into a night scene, apply Van Gogh-style brush strokes, or generate entirely new compositions based on the original image content. All processing happens entirely on-device on supported Pixel devices using the Tensor NPU, ensuring privacy and speed. Reimagine was introduced with the Pixel 9 series as an experimental creative tool.
Everyday Use Cases
- Artistic photo transformation
- Style transfer
- Lighting changes
- Color palette shifts
- Composition reinterpretation
- Creative experimentation
User Workflow
- User selects a photo in Reimagine
- Chooses a style or describes a transformation in text
- Gemini Nano analyzes the original image content
- AI generates artistic interpretations on-device
- User selects and saves the favorite result
AI Processing Flow
The source photo is analyzed by on-device Gemini Nano vision models that detect objects, scenes, and composition. A generative diffusion model transforms the image based on the selected style or text prompt. Neural style transfer networks apply artistic effects. All processing uses the Tensor NPU with no data leaving the device. Multiple variants are generated for user selection.
Inputs / Outputs
Inputs:
- Source photo
- Style selection or text prompt
- Color preferences
- Composition guidance
- Lighting style
- Artistic reference
Outputs:
- Stylized photo variants
- Reimagined compositions
- Color-transformed images
- Lighting-changed versions
- Artistic interpretations
Known Limitations
- Only available on Pixel 9 series and newer
- Quality varies by selected style
- May introduce artifacts with complex source images
- Generation takes 15 to 30 seconds
- Limited artistic styles compared to cloud AI
Unsupported Scenarios
- Realistic face swaps
- Professional art reproduction
- Commercial use
- Non-Pixel devices
- Video frame transformation
Performance Notes
Image analysis in under 1 second on Tensor G4; Style transformation in 15 to 25 seconds; NPU acceleration for generative model; Battery consumption 10 to 15 percent per session
Available On
Shows where this feature is available and how its AI processing works on each platform. Availability may vary by device.
| Platform | Execution | Offline | Cloud | OS / Software |
|---|---|---|---|---|
| Gemini Nano | Local | Yes | No - on-device | Android 15+ |
Research Status
Confidence and verification reflect how complete documentation is. Fields may show Not assessed or Not yet verified while research is ongoing - this flags gaps, not product deficiencies.
| Research Status | Verified |
| Confidence | High |
| First introduced | 2025-08-13 |
| Last updated | 2026-08-16 |
| Last verified | 2026-08-16 |
Research Notes
Available on Pixel 9 and newer with Tensor G4; Introduced October 2024 with Pixel 9; Uses Gemini Nano 2 generative model; Experimental feature with regular style updates; 12 artistic styles at launch
Details
| Vendor | |
| AI Platform | Gemini Nano |
| Category | Image Generation |
Capability Mapping
This feature maps to canonical capability family: Image Generation.
Canonical capability: Image Generation.
- Image Generation - Creates novel imagery or emoji-like generated outputs from prompts.
Grounded in 4 device(s) on Gemini Nano.
Related Features
- Photo Assist - Cross-platform equivalent
This page describes platform-level capability; not device-level support for any specific product.