Kie
Integrates Gemini image generation models into apps via a unified API for text-to-image and multi-turn editing
Last verified:
What is Kie?
Nano Banana Pro API is a Google‑powered visual AI service delivered through Kie.ai that lets developers integrate Gemini 3 Pro Image and Gemini 2.5 Flash Image models into their apps, websites, and workflows. It provides a unified API for both the speed‑focused Nano Banana API (Gemini 2.5 Flash) and the higher‑fidelity Nano Banana Pro API (Gemini 3 Pro Image), enabling everything from rapid low‑cost edits to 4K studio‑quality image generation. The tool supports text‑to‑image, image‑to‑image editing, multi‑image composition, and search‑driven visuals, with fine‑grained control over resolution, aspect ratio, output format, and NSFW checking.
Key features include multi‑turn editing where you refine images over several API calls instead of regenerating from scratch, support for up to 8 reference images that are blended into one coherent output, and long‑context prompts up to 64K input and 32K output tokens for complex creative briefs. The Nano Banana Pro API excels at photorealistic renders, accurate text‑in‑image (logos, posters, UI mockups, packaging), and consistent character identity across scenes for storyboards or brand‑kit assets. It also integrates search‑grounded knowledge from Google so diagrams, infographics, and technical visuals can reflect up‑to‑date or domain‑specific references.
This tool is aimed at developers, AI‑tool builders, SaaS platforms, and product teams who want an affordable, API‑first way to embed Gemini‑based image generation and editing into their products. Typical use cases include creative apps with instant previews, e‑commerce mockups, marketing and ad‑creation tools, design systems, and batch‑processing pipelines for large image libraries. The flexible credit‑based pricing and unified documentation make it suitable both for prototyping and large‑scale production workloads that need stable, high‑concurrency performance.
Kie pricing
Pricing model: Freemium
Nano Banana API pricing starts at about $0.02 per image via Kie.ai’s credit‑based system, billed per generation or edit with no mandatory subscriptions, making it suitable for lightweight editing and large‑scale batch workflows. Nano Banana Pro API pricing begins at around 24 credits per image (approximately $0.12), which is roughly 20% cheaper than the official Google price for Gemini 3 Pro Image at launch, and supports 4K output, multi‑turn workflows, and long‑context prompts. Kie.ai uses a flexible pay‑as‑you‑go credit model where users buy credits and consume them across both Nano Banana and Nano Banana Pro without separate subscriptions, with optional bulk credits for scaling to production‑level workloads.
Kie pros
- Unified API key for both Nano Banana and Nano Banana Pro
- Seamless switching between Gemini 2.5 Flash and Gemini 3 Pro Image
- Fast generation speeds on Nano Banana API (Gemini 2.5 Flash)
- Low cost per image on Nano Banana API (~$0.02 per image)
- 4K high‑resolution output on Nano Banana Pro API
- 64K input and 32K output context window for long creative briefs
- Multi‑turn editing without full regeneration
- Up to 8 reference images blended into one output
- Precise text‑in‑image rendering for posters and UI mockups
- Studio‑quality control over lighting, composition, and depth of field
- Strong character and object consistency across scenes
- Search‑driven accuracy for diagrams and infographics
- Multi‑language text and localization inside visuals
- Handles complex, multi‑step workflows and conditional instructions
- Realistic physics‑aware lighting, shadows, and spatial relationships
- Flexible pay‑as‑you‑go credits with no mandatory subscriptions
- Stable performance and high‑concurrency support
- Unified documentation and simple model switching
- Built‑in NSFW checker parameter
- Supports common image formats (JPEG, PNG, WEBP) and multiple file uploads
Kie cons
- Nano Banana Pro is slower than the standard Nano Banana API
- Higher cost per image on Nano Banana Pro (~$0.12 per image)
- Pricing subject to change as models evolve
- Limited multi‑image workflows on the base Nano Banana API
- Not all advanced features available on the 2.5 Flash tier
- Requires API integration work for non‑technical users
- No native GUI builder or visual editor on the Kie site
- Users must manage credits and spending outside their own systems
- Potential latency for very large or complex prompts
- No guarantee of perfect text rendering in all edge cases
Frequently asked questions about Kie
What is the difference between Nano Banana API and Nano Banana Pro API?
Nano Banana API is built on Gemini 2.5 Flash Image and focuses on fast, low‑cost image edits and generation, ideal for high‑volume or real‑time applications. Nano Banana Pro API uses Gemini 3 Pro Image and adds 4K resolution, long‑context prompts (64K input, 32K output), multi‑turn editing, and advanced control over text, layout, and complex scenes, targeting professional creators and enterprises needing studio‑quality assets.
Do I need separate API keys for Nano Banana and Nano Banana Pro?
No, Kie.ai provides a unified API key that grants access to both Nano Banana API (Gemini 2.5) and Nano Banana Pro API (Gemini 3 Pro Image) through the same authentication flow, so you can switch between models without managing multiple keys.
How many images can I edit or reference at once?
The Nano Banana Pro API supports blending up to 8 reference images into one coherent output, aligning lighting, perspective, and style across them. The standard Nano Banana API allows multiple image edits but is more limited in multi‑image composition workflows compared to the Pro tier.
What resolutions and aspect ratios are available?
Nano Banana API supports configurable image sizes such as 1:1, 9:16, 16:9, 3:4, 4:3, 3:2, 2:3, 5:4, 4:5, 21:9, and auto. Nano Banana Pro API adds resolution options of 1K, 2K, and 4K plus the same set of aspect ratios, allowing outputs suitable for printing, packaging, and high‑detail designs.
Can the APIs generate text inside images accurately?
Nano Banana Pro API delivers sharp, legible text built into images, making it suitable for posters, UI mockups, product packaging, and technical diagrams with complex wording. It improves on common glitchy‑text issues in basic models and handles longer prompts with accurate spelling and multilingual support, while the standard Nano Banana API is more limited in this regard.
How does multi‑turn editing work?
Nano Banana Pro API allows you to refine an image over several API calls instead of regenerating from scratch, maintaining context so you can adjust layout, lighting, or typography across iterations. Gemini 3 Pro Image keeps the prior state and edits consistent, letting you treat image generation as a dialogue‑style workflow rather than one‑off prompts.
What kind of NSFW or safety controls are available?
Both tiers expose an nsfw_checker parameter that can be toggled to enable or disable NSFW detection, letting you control whether explicit content is filtered at the API level. This helps developers build safer creative tools that comply with their own content policies and platform requirements.
How does pricing compare to Google’s own Gemini Image pricing?
Kie.ai positions Nano Banana Pro API at about 20% cheaper than the official Google Gemini 3 Pro Image price at launch, with typical Nano Banana Pro images billed at roughly $0.12 per image via the credit system. Nano Banana API starts around $0.02 per image, also significantly undercutting many mainstream image‑generation platforms for large‑volume editing tasks.
Can I use this API for batch processing large image libraries?
Yes, Kie.ai optimizes both Nano Banana and Nano Banana Pro APIs for high‑concurrency workloads and batch processing, offering stable performance and predictable response times under heavy loads. The low‑cost Nano Banana API is especially suited for large‑scale edits or instant preview pipelines, while the Pro tier fits premium, high‑fidelity batches where quality matters more than speed.
What formats and file limits apply to uploads?
For image‑to‑image and editing endpoints, supported formats are JPEG, PNG, and WEBP with maximum file sizes up to 10 MB for editing and 30 MB for generation, and up to 8 input images per request for transformation or reference. The APIs also support providing image URLs via image_urls or image_input arrays, giving flexibility between direct uploads and remote references.