Quick Start

From zero to your first AI-generated image in under 2 minutes.

Install the Addon

1. Download nano_blender_v1.1.0.zip
2. Open Blender and go to Edit > Preferences > Add-ons
3. Click Install... (top right), select the zip file
4. Enable "Nano Blender" by checking the box

Install Dependencies

1. Press N in the 3D Viewport to open the sidebar
2. Click the Nano Blender tab
3. Click Install Dependencies (installs the google-genai package)

Get Your API Key

1. Go to aistudio.google.com/apikey (or click Get API Key in the setup panel)
2. Sign in with your Google account
3. Click Create API Key and copy it

Configure the Addon

1. Go to Edit > Preferences > Add-ons, find "Nano Blender" and expand it
2. Paste your API key in the Gemini API Key field
3. Click Validate API Key
4. You should see "Flash tier (free features)" or "Pro tier (all features unlocked)"

Generate Your First Image

1. Switch to the Image Editor workspace (or split a window and set it to Image Editor)
2. Press N to open the sidebar, click the Nano Blender tab
3. In the Generate panel, type a prompt:
   A cozy mountain cabin at sunset, photorealistic, warm lighting
4. Choose an aspect ratio (or leave on Auto)
5. Click Generate Image
6. Wait a few seconds — your image appears in the Image Editor

That's it. You're up and running.

Troubleshooting

"Install Dependencies" fails

• Make sure Blender has internet access
• Try running Blender as administrator (Windows) or with sudo (Linux)
• Check if your firewall is blocking Python/pip

"Invalid API Key"

• Make sure you copied the full key (no extra spaces)
• Generate a new key at aistudio.google.com
• Keys look like: AIzaSy... (39 characters)

"No image returned"

• Google may have blocked the generation due to safety filters
• Try rephrasing your prompt

Panels not showing in Image Editor

• Press N to open the sidebar
• Look for the Nano Blender tab
• Make sure your API key is entered in addon preferences

Next: All Features →

All Features

A detailed guide to every feature in Nano Blender.

1. Text-to-Image Generation Free

Location: Image Editor > Sidebar > Nano Blender > Generate

Creates an image from a text description. The result loads directly into the Image Editor.

How to use

1. Open the Image Editor and press N to open the sidebar
2. Select Flash or Pro model (Pro requires billing)
3. Type your prompt in the text field
4. Optionally select an aspect ratio (1:1, 16:9, 9:16, etc.)
5. Click Generate Image

Tips

• Be descriptive: "A medieval blacksmith's workshop, dimly lit by forge fire, iron tools on stone walls, photorealistic" works better than "blacksmith shop"
• Specify camera and lighting for photorealism: "shot with a 35mm lens, soft natural lighting"
• For textures, mention "seamless tileable texture" in your prompt

2. Image Editing Free

Location: Image Editor > Sidebar > Nano Blender > Edit Image

Modifies the currently open image based on text instructions. The panel automatically picks up whatever image is displayed in the Image Editor — no manual selection needed.

How to use

1. Open an image in the Image Editor (or generate one first)
2. The panel shows "Editing: [image name]" confirming which image will be edited
3. Describe what you want to change
4. Click Edit Image

Example prompts

• "Change the sky to a dramatic sunset"
• "Remove the person in the background"
• "Add snow to the ground and rooftops"
• "Make the lighting warmer and more golden"
• "Replace the wooden door with a metal one"

3. AI Rendering Free

Location: 3D Viewport > Sidebar > Nano Blender > AI Render

The signature feature. Build a rough scene with basic shapes and solid colors, then let AI transform it into a finished image. See the dedicated AI Rendering guide for a full tutorial.

How to use

1. Build a simple scene (cubes, cylinders, basic colored materials)
2. Choose source: Viewport (faster) or Camera (uses render settings)
3. Write a prompt describing the final look
4. Click AI Render

4. Render Post-Processing Free

Location: 3D Viewport > Sidebar > Nano Blender > AI Render (bottom section)

Enhances or stylize your finished Blender render with a text prompt.

How to use

1. Render your scene normally (F12)
2. Type an enhancement prompt: "make it cinematic and moody"
3. Click Process Current Render

Auto mode: Toggle Auto Post-Process to automatically process every render as soon as it finishes.

Example prompts

• "Film still, shallow depth of field, cinematic color grading"
• "Add atmospheric fog and volumetric lighting"
• "Watercolor painting style"
• "Warm vintage tone, subtle grain"

5. Style Transfer Free

Location: 3D Viewport > Sidebar > Nano Blender > Style Transfer

Takes the artistic style of one image and applies it to another.

How to use

1. Select a Content Image (the image you want to restyle)
2. Select a Style Image (the image whose style you want to copy)
3. Optionally adjust the prompt (the default works well)
4. Click Style Transfer

Example uses

• Apply a Van Gogh painting style to your 3D render
• Take a concept art style and apply it to a blockout render
• Match the color palette and mood of a reference photo

Pro Features

The following features require a Gemini billing account. See the Pro Features guide for full details.

AI Rendering

The signature feature. Turn rough 3D blockouts into finished images in seconds.

The Concept

Tutorial: Medieval Village

Goal: Create a medieval village image from scratch in 10 minutes.

Step 1 — Block out the ground

Add a plane and scale it to be your ground. Assign a green material (grass hint).

Step 2 — Block out buildings

Add cubes for building bodies, scale them to different sizes. Add smaller cubes or triangular shapes on top for roofs. Assign tan/brown materials to walls, dark brown/red to roofs.

Step 3 — Add environment

• Tall thin cylinders with green spheres on top = trees
• A flat blue plane = a pond or river
• Small cubes along the ground = rocks or crates

Step 4 — Position your view

Move to a nice camera angle. Choose Viewport as source for quick results, or set up a camera and choose Camera for specific framing.

Step 5 — Write your prompt

Photorealistic medieval village, thatched roof cottages, cobblestone paths,
lush green countryside, warm afternoon sunlight filtering through trees,
volumetric fog in the distance, highly detailed, 8K quality

Step 6 — Click AI Render

Wait 5–10 seconds. The result appears in the Image Editor.

Step 7 — Iterate

Adjust your blockout (move buildings, change colors), tweak the prompt, render again.

Color Coding Guide

Use material colors as semantic hints for the AI:

Prompt Formula

Structure your AI Render prompts like this:

[Subject/scene type], [key details], [material/texture descriptions],
[lighting], [mood/atmosphere], [quality keywords]

Examples

Modern minimalist house, white concrete walls, floor-to-ceiling windows,
surrounded by a Japanese garden, soft overcast lighting, architectural
photography, clean lines

Ancient elven forest temple, overgrown with moss and vines, crystal
clear stream, magical glowing runes, ethereal misty atmosphere,
concept art style, highly detailed

Cyberpunk city street at night, neon signs reflecting on wet asphalt,
hovering vehicles, dense urban environment, volumetric fog,
cinematic wide angle shot

Common Mistakes

Viewport vs Camera

Prompt Tips

Write better prompts. Get better images.

General Principles

1. Be specific, not vague. "A red sports car" produces generic results. "A cherry red Ferrari F40 on a coastal highway at golden hour, shot from a low angle with a 24mm lens" produces memorable ones.
2. Describe the scene narratively. Think of it as telling someone what to paint, not listing keywords.
3. Mention the medium/style. "Photorealistic", "watercolor painting", "concept art", "oil painting", "pencil sketch" — this one word changes everything.
4. Include lighting. Lighting makes or breaks an image. "Warm golden hour light", "harsh overhead noon sun", "soft overcast", "dramatic rim lighting".
5. Specify camera when going photorealistic. "Shot with a 35mm lens", "shallow depth of field", "wide angle", "aerial drone shot".

Prompt Templates

[Subject], [environment details], [material descriptions], shot with a
[focal length] lens, [lighting type], [time of day], photorealistic,
highly detailed, [mood]

[Subject] in [setting], [art style] style, [color palette],
[atmospheric effects], concept art, detailed illustration,
[artist reference if applicable]

[Product] on [surface], [background], [lighting setup], commercial
product photography, [camera angle], sharp focus, [material details]

Seamless tileable [material] texture, [color/variation details],
[surface quality], [weathering/age], top-down view,
uniform lighting, PBR ready

[Character description], [pose], [clothing/armor details],
[expression], [background], [art style], full body/portrait,
detailed, [lighting]

What Works Well

• Camera specs: "85mm portrait lens", "macro lens", "fisheye"
• Specific art styles: "Studio Ghibli style", "minimalist flat design"
• Material details: "brushed aluminum", "worn leather", "frosted glass"
• Atmosphere: "foggy", "dusty", "misty rain", "snow flurries"
• Time references: "1920s art deco", "futuristic year 3000", "ancient Roman"
• Emotional tone: "serene", "ominous", "joyful", "melancholic"

What to Avoid

• Conflicting instructions: "dark and bright" confuses the model
• Too many subjects: Focus on one main subject, not ten
• Negative language: Instead of "no people", describe what IS there
• Extremely long prompts: 2–4 sentences is the sweet spot
• Exact text: Gemini can render text but it's hit-or-miss. Keep text requests short

Feature-Specific Tips

For Image Editing

• Be explicit: "Change the sky to sunset" not "change the sky"
• Reference positions: "the object on the left", "the background"
• Preserve context: "Keep everything else the same, only change..."

For Style Transfer

• The default prompt works well. Only customize if you want specific blend behavior.
• "Apply the color palette and brush strokes of the style image while keeping the composition of the content image"

For AI Rendering

• Describe the FINAL look, not the blockout
• Focus on mood, materials, and lighting
• Let the AI handle detail — your blockout provides structure

For Post-Processing

• Short prompts work great: "cinematic", "warm and cozy", "dark and moody"
• Combine style + mood: "Wes Anderson color palette with soft vintage grain"

Pro Features

Unlock the full power of Nano Blender with a Gemini billing account.

How to Enable Pro

1. Go to aistudio.google.com
2. Sign in with the same Google account you used for your API key
3. Navigate to Settings > Billing and enable billing with a payment method
4. Back in Blender, click Validate API Key again
5. Status should update to "Pro tier (all features unlocked)"

Hi-Res Generation (2K/4K) Pro

Location: 3D Viewport > Sidebar > Pro Features

Standard Flash generation produces images around 1024px. Hi-Res mode generates at 2K (2048px) or 4K (4096px).

When to use

• Final output for portfolio pieces
• Print-quality images
• Large format textures
• Detailed scenes where you need to zoom in

Recommended workflow

1. Prototype with standard generation (fast, free)
2. When you have the right prompt, switch to Hi-Res
3. Select 2K or 4K resolution
4. Generate the final version

Multi-Reference Generation Pro

Location: 3D Viewport > Sidebar > Pro Features

Feed up to 14 reference images to guide the AI. The result blends elements from all references with your text prompt.

When to use

• Maintaining consistent style across a project
• "More like these" generation
• Architectural style matching
• Character consistency (approximate)

Example

References: 5 photos of Art Deco buildings
Prompt: "A new Art Deco skyscraper at night, matching the architectural style of the reference images, dramatic uplighting"

Search-Grounded Generation Pro

Location: 3D Viewport > Sidebar > Pro Features

Generates images incorporating real-time information from Google Search. The AI searches the web before generating, so results reflect current data.

Example prompts

• "The latest Tesla Cybertruck in a desert landscape at sunset"
• "Visualization of today's global temperature map"
• "A street scene in Tokyo's Shibuya district, current architecture"

Multi-Turn Editing Pro

Location: 3D Viewport > Sidebar > Pro Features

A chat-style session where each message builds on the previous result. The AI remembers the full conversation.

Example session

1. Click Start Session
2. "Create a cozy coffee shop interior, morning light through windows"
3. See result. "Add more books on the shelves and a cat sleeping on the counter"
4. See result. "Change the lighting to evening, warm lamp light"
5. See result. "Add rain on the windows"
6. Perfect. Click End Session

Tips

• Keep instructions short and focused on one change per turn
• You can undo by saying "go back to before the last change"
• Sessions are not saved between Blender restarts
• Start a new session for a completely different image

Frequently Asked Questions

Quick answers to the most common questions.

Setup & Installation

How do I install Nano Blender?

Download the zip file. In Blender, go to Edit > Preferences > Add-ons > Install, select the zip, and enable "Nano Blender". See the Quick Start guide for full steps.

What does "Install Dependencies" do?

It installs the google-genai Python package into Blender's bundled Python. This is the Google SDK needed to communicate with Gemini's API. It's a one-time installation.

The dependency installation failed. What do I do?

Try running Blender as administrator (Windows: right-click > Run as administrator). If behind a corporate firewall, it may block pip. You can also manually install via Blender's Python console:

import subprocess, sys
subprocess.check_call([sys.executable, '-m', 'pip', 'install', 'google-genai'])

Where do I get an API key?

Go to aistudio.google.com/apikey, sign in with your Google account, and click "Create API Key". It's free and takes 30 seconds.

Is the API key free?

Yes. Google provides a free tier for Gemini with generous rate limits. You can generate hundreds of images per day at zero cost.

Features & Usage

Do I need a powerful GPU?

No. All AI processing happens on Google's cloud servers. Nano Blender works on any machine that runs Blender 4.0+, including laptops without dedicated GPUs.

How long does generation take?

Typically 3–10 seconds. First generation after launching Blender may take slightly longer (cold start). Hi-res and Pro features take a bit longer than standard Flash.

Can I use generated images commercially?

Images are generated by Google Gemini. Google's terms of service for AI-generated content apply. Check the current policy for your specific use case at ai.google.dev/terms.

What image resolution do I get?

Flash model generates around 1024px on the longest side. Pro Hi-Res mode generates at 2K (2048px) or 4K (4096px).

Does it work offline?

No. An internet connection is required to reach Google's Gemini API.

Where are generated images saved?

By default, images save to a nano_blender_output/ folder next to your .blend file. You can change the folder name or disable auto-save in addon preferences.

Can I use generated images as textures?

Yes. Generated images load into bpy.data.images. You can use them in shader nodes, UV project them, or apply them as textures just like any other image in Blender.

Free vs Pro

What's the difference between Free and Pro?

Free tier (Flash) gives you 5 features: text-to-image, image editing, AI rendering, style transfer, and render post-processing. Pro adds 4 more: hi-res 2K/4K, multi-reference (14 images), search-grounded generation, and multi-turn editing. Pro also gives higher quality on the standard Generate feature via the model selector.

How much does Pro cost?

Pro requires enabling billing on your Google AI account. API costs are roughly $0.01–0.05 per generation. Most users spend $1–5/month.

Do I need Pro to use AI Rendering?

No. AI Rendering works on the free Flash tier.

Comparison

How is this different from Dream Textures / ComfyUI addons?

Those use Stable Diffusion (local or cloud), requiring a local GPU, model downloads (4–8GB+), and complex setup. Nano Blender uses Google Gemini via API — no local GPU, no downloads, no extra software. Paste a key and go.

Can I use this alongside other AI addons?

Absolutely. Nano Blender is complementary. Use it when you want results in 10 seconds without node setup.

Troubleshooting

I get "No image returned" after generating.

Google's safety filters may have blocked the generation. Try rephrasing your prompt. Also ensure your API key is valid (re-validate in preferences).

The addon panel doesn't show up.

Make sure the addon is enabled in Preferences. Generate and Edit panels are in the Image Editor sidebar (N). Other panels (AI Render, Style Transfer, Pro Features) are in the 3D Viewport sidebar.

Generation seems stuck or very slow.

Check your internet connection. The first call may be slower due to dependency loading. If consistently slow, Google's API may be under high load — try again in a few minutes.

I'm getting rate limit errors.

The free tier has rate limits. Wait a minute and try again. Pro tier has higher limits.

Support

What if I don't like it?

Contact us within 14 days of purchase for a full refund. No questions asked.

Will the addon be updated?

Yes. All updates are included with your purchase. We update for new Blender versions and new Gemini model releases.

What if Google changes their API?

We'll update the addon. Since you bring your own API key, you're never locked into our pricing — you pay Google directly.

Changelog

All notable changes to Nano Blender.

v1.1.0 — 2026-02-17

Changed

• Generate and Edit Image panels moved to the Image Editor sidebar for a more natural workflow
• Edit Image now auto-picks the currently open image — no manual dropdown needed
• Added Flash / Pro model selector to the Generate panel. Pro is greyed out when billing is not enabled.
• When billing is detected, the addon defaults to the Pro model for higher quality generation
• API key validation now uses a free metadata call instead of a full generation — essentially zero cost

Removed

• Removed the Compose Images feature (rarely used, confusing UX)

v1.0.0 — 2026-02-17

Initial Release

Flash Features (Free Tier):

• Text-to-image generation with aspect ratio control
• Image editing with text instructions
• AI Rendering (blockout to finished image)
• Render post-processing with auto-process option
• Style transfer between two images

Pro Features (Billing Tier):

• Hi-res 2K/4K generation
• Multi-reference generation (up to 14 images)
• Search-grounded generation with live Google Search data
• Multi-turn iterative editing sessions

Infrastructure:

• Auto-dependency installation (google-genai)
• API key validation with automatic tier detection
• Background processing (non-blocking UI)
• Auto-save to output folder
• Blender 4.0+ compatibility