Two Fal endpoints: `fal-ai/nano-banana` (text-to-image) and `fal-ai/nano-banana/edit` (image-to-image).
Fal.ai docs← Back to modelsModes Inputs Outputs Pricing Seed & batches
Nano Banana – Photoreal Text-to-Image + Image Editing
Image + Edit
Nano Banana ships with two Fal endpoints: lightning-fast text-to-image and precise image-to-image editing. It shares the same wallet, logs, and prompt lab as our video engines—so you can storyboard frames, iterate on references, or prep shot plates directly inside MaxVideoAI.
Where Nano Banana shines
- Photoreal concept frames and hero stills
- Variant exploration before animating in Veo / Sora
- High-quality marketing art without leaving MaxVideoAI
- Precise image edits (prop swaps, wardrobe tweaks, room re-lighting)
Technical overview
`prompt` (required), `num_images` (1–8). Edit mode adds `image_urls` for source uploads.
High-resolution JPEG/PNG URLs with metadata (description + CDN URL). Each run is logged in Recent Runs & Library.
Live wallet-based pricing — open the Generate workspace for current rates.
Lock the seed to reproduce looks; use `num_images` for multi-shot batches.
Prompt mini-guide
“Editorial fashion portrait of a woman in a chrome raincoat under neon signage, 85mm lens, shallow depth of field, cinematic grading.”
Keep prompts descriptive but grounded:
- Anchor subject + setting (“portrait of… inside a Tokyo jazz bar”).
- Add lens and lighting cues for style (film emulation, aperture, color palette).
- Mention mood / action to avoid generic poses.
- For edits, describe the transformation (“replace the jacket with holographic fabric…”).
- Use `num_images` for variety; reuse the seed when you find a favorite.
Tips & tricks
- 🖼 Generate keyframes, then feed them into Veo/Sora shot-by-shot.
- 🎨 Treat edit mode as a “texture re-roll” for props, lighting, or wardrobe.
- 📚 Organize outputs via Library → Images to keep campaigns tidy.
- ⚡ Use `num_images = 4` with batch naming to compare variations quickly.
- 🔁 Pair with the new Image Jobs view to audit spend per concept.
Compare Nano Banana with AuraFlow, Flux et al. in the guide moteurs.
Overview
- Brand
- Engine ID
- nano-banana
- Slug
- /models/nano-banana
- Logo policy
- Text-only (wordmark)
- Live pricing updates inside the Generate workspace.
- Live pricing updates inside the Generate workspace.
Live pricing updates inside the Generate workspace. Check the current rate before you run a job.
Prompt ideas
Lifestyle hero shot
Golden hour photo of a runner splashing through a Los Angeles backyard pool, high-speed shutter, crisp water droplets, Leica 50mm.
Edit prompt
Make the living room brighter, swap the coffee table for a glass slab, add a neon “BANANA” sign on the back wall.
FAQ
How many images per run?
Set `num_images` between 1 and 8. Each image is logged separately, so you can download / favorite individually.
Does edit mode require masks?
No—describe the transformation in plain language. For precise regions, upload cropped inputs or run multiple passes.
Where do the images live?
Every output hits Recent Runs + the new Library → Images tab. You can copy the CDN link or download lossless files.
Explore other engines
Compare price tiers, latency, and prompt presets across the rest of the catalog.
google-veo
Google Veo 3.1
Generate cinematic 8-second videos with native audio using Veo 3.1 by Google DeepMind on MaxVideoAI. Reference-to-video guidance, multi-image fidelity, pay-as-you-go pricing from $0.52/s.
Try Google Veo 3.1google-veo
Google Veo 3.1 Fast
Use Veo 3.1 Fast for affordable, fast AI video generation. Up to 8-second clips with optional native audio—ideal for social formats and iterative testing.
Try Google Veo 3.1 Fastgoogle-veo
Google Veo 3.1 First/Last Frame
Upload starting and ending frames, write a brief, and let Veo 3.1 animate seamless transitions with optional native audio. Swap to Fast mode for cheaper iterations.
Try Google Veo 3.1 First/Last Frame