Nano Banana Pro and Sora-2 Integration with ComfyUI Complete Configuration Guide: Simplest API Integration Solution

As an open-source image and video generation tool, ComfyUI has become the go-to platform for AI creators. With the release of Google's Nano Banana Pro image model and OpenAI's Sora-2 video model, many users are asking a key question: Can these cutting-edge models be integrated into ComfyUI? And what's the simplest way to configure them?

This article dives into these questions, providing a complete integration guide ranging from official nodes to third-party API platforms. We'll help you quickly leverage the power of Nano Banana Pro and Sora-2 within your ComfyUI workflows.

1. Three Ways to Connect Nano Banana Pro API to ComfyUI

Nano Banana Pro is a flagship image generation model released by Google DeepMind, supporting 4K resolution, multilingual text rendering, and high-fidelity character consistency. Based on official documentation and community practices, there are three main ways to connect it to ComfyUI:

1.1 Official ComfyUI Partner Nodes (Recommended for Beginners)

This is the easiest integration method and doesn't require installing any extra plugins. ComfyUI has already integrated Nano Banana Pro as native API nodes.

Steps:

# Step 1: Update ComfyUI to the nightly version
cd ComfyUI
git pull origin main
python main.py

# Step 2: Inside the ComfyUI interface:
# 1. Double-click the empty canvas
# 2. Search for the "Google Gemini Image" node
# 3. Drag it into your workflow

Node Configuration Parameters:

Pros: Works out of the box; no server or proxy configuration needed.
Cons: Requires a Google AI Studio API key, which might be restricted in some regions.

🎯 Pro-tip: For individual users and small-scale testing, we recommend the official Partner Nodes. If you need more stable, enterprise-grade service, you can get high-availability API access through the APIYI (apiyi.com) platform. It uses the native Google format, so you can swap out the official API without changing any code while reducing costs by 80%.

1.2 Custom Nodes (Best for Advanced Users)

The GitHub community offers an open-source "ComfyUI-NanoBanano" custom node set that supports more advanced features.

Installation Steps:

# Navigate to the ComfyUI custom nodes directory
cd ComfyUI/custom_nodes

# Clone the repository
git clone https://github.com/ShmuelRonen/ComfyUI-NanoBanano.git

# Install dependencies
cd ComfyUI-NanoBanano
pip install -r requirements.txt

# Restart ComfyUI

Supported Operation Modes:

Pros: More feature-rich; supports multimodal operations.
Cons: Requires manual maintenance and plugin updates.

1.3 Third-party API Platforms (Enterprise-grade Solution)

For enterprise users requiring high concurrency, low latency, and cost optimization, third-party API platforms offer a better way forward.

Core Advantage Comparison:

Configuration Example (using APIYI as an example):

# Original Google Official Config
api_endpoint = "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp:generateImage"
api_key = "YOUR_GOOGLE_API_KEY"

# Switching to the APIYI platform (only change these two lines)
api_endpoint = "https://api.apiyi.com/v1/models/gemini-2.0-flash-exp:generateImage"
api_key = "YOUR_APIYI_API_KEY"  # Get this from apiyi.com

🎯 Enterprise Recommendation: If your team needs to process image generation tasks in bulk within ComfyUI, we recommend using the APIYI (apiyi.com) platform to access the Nano Banana Pro API. It provides dedicated acceleration channels and a pay-as-you-go model, making it ideal for production environments processing over 1,000 images per day.

2. Sora-2 Video Model Integration in ComfyUI: Feasibility Analysis and Configuration

Addressing the core question from users: "Is it possible to use this video model in comfyUI?" The answer is yes, but you'll need to access it via API nodes.

2.1 Technical Principles of Sora-2 Integration with ComfyUI

Sora-2 is the second-generation video generation model released by OpenAI. Unlike image models, Sora-2 cannot be deployed locally like Stable Diffusion. Accessing it within ComfyUI requires one of the following two paths:

Conclusion: Sora-2 must be accessed via API nodes in ComfyUI; local model deployment isn't supported.

2.2 Official Sora-2 API Nodes Configuration Tutorial

ComfyUI has officially released Sora-2 Partner Nodes, supporting both text-to-video (t2v) and image-to-video (i2v) modes.

Installation Steps:

# Step 1: Ensure ComfyUI version ≥ v0.2.6
cd ComfyUI
git pull origin main

# Step 2: Add Sora-2 nodes to your workflow
# 1. Double-click the empty canvas
# 2. Search for "OpenAI Sora" or "Sora-2-i2v"
# 3. Connect input nodes (text prompt or reference image)

Sora-2 Pro Node Parameter Configuration:

Important Notes:

• ⚠️ Resolution Matching: In image-to-video mode, your input image must exactly match the selected video resolution, or you'll get an error.
• ⚠️ Cost Control: Sora-2 Pro is billed by the second; a 20-second video costs roughly $0.50-$1.00 USD.
• ⚠️ API Quota: OpenAI has daily call limits for new accounts; you'll need to apply for a quota increase in advance.

{{SVG_TEXT_0}}: Sora-2 Cost Comparison Analysis
{{SVG_TEXT_1}}: Estimated Costs by Resolution and Duration (USD)
{{SVG_TEXT_2}}: Configuration Plan
{{SVG_TEXT_3}}: Resolution
{{SVG_TEXT_4}}: Duration
{{SVG_TEXT_5}}: Official API
{{SVG_TEXT_6}}: APIYI Platform
{{SVG_TEXT_7}}: Sora-2 Standard
{{SVG_TEXT_8}}: Save 40%
{{SVG_TEXT_9}}: Sora-2 Standard
{{SVG_TEXT_10}}: Save 40%
{{SVG_TEXT_11}}: Sora-2 Pro
{{SVG_TEXT_12}}: Save 37.5%
{{SVG_TEXT_13}}: Sora-2 Pro
{{SVG_TEXT_14}}: Save 37.5%
{{SVG_TEXT_15}}: 💰 Cost Optimization Strategies
{{SVG_TEXT_16}}: • Shorten video duration (20s→10s) = 50% savings
{{SVG_TEXT_17}}: • Lower resolution (1792→1280) = 30% savings
{{SVG_TEXT_18}}: • Use APIYI Platform = Extra 35-40% savings
{{SVG_TEXT_19}}: 🎯 Enterprise Plan Advantages
{{SVG_TEXT_20}}: • Tiered Pricing: $500+/mo consumption enjoys a 45% discount
{{SVG_TEXT_21}}: • Dedicated Channel: High concurrency and no rate limits guaranteed
{{SVG_TEXT_22}}: • Tech Support: 24/7 support from dedicated engineers
{{SVG_TEXT_23}}: Visit apiyi.com to register; new users get a $20 free credit | Supports unified access for Nano Banana Pro + Sora-2

{{SVG_TEXT_0}} {{SVG_TEXT_1}}

{{SVG_TEXT_2}} {{SVG_TEXT_3}} {{SVG_TEXT_4}} {{SVG_TEXT_5}} {{SVG_TEXT_6}}

{{SVG_TEXT_7}} 1280×720 5s $0.25 $0.15 {{SVG_TEXT_8}}

{{SVG_TEXT_9}} 1280×720 10s $0.50 $0.30 {{SVG_TEXT_10}}

{{SVG_TEXT_11}} 1792×1024 10s $0.80 $0.50 {{SVG_TEXT_12}}

{{SVG_TEXT_13}} 1792×1024 20s $1.60 $1.00 {{SVG_TEXT_14}}

{{SVG_TEXT_15}} {{SVG_TEXT_16}} {{SVG_TEXT_17}} {{SVG_TEXT_18}}

{{SVG_TEXT_19}} {{SVG_TEXT_20}} {{SVG_TEXT_21}} {{SVG_TEXT_22}}

{{SVG_TEXT_23}}

2.3 Third-Party Sora-2 API Nodes (Siray Solution)

In addition to the official nodes, the community provides an enhanced version called Sora-2 Custom Nodes.

Siray OpenAI Sora-2-i2v Node Features:

# Install Siray custom nodes
cd ComfyUI/custom_nodes
git clone https://github.com/siray-comfyui/openai-sora-2.git
pip install -r openai-sora-2/requirements.txt

Enhanced Features:

🎯 Workflow Advice: For teams needing to generate large volumes of marketing videos, we recommend building an automated workflow by combining ComfyUI's preprocessing nodes with the Sora-2 API. You can get a more stable Sora-2 API interface through the APIYI (apiyi.com) platform. It's especially suited for batch video generation, offering dedicated queue management and cost optimization solutions.

3. Nano Banana Pro and Sora-2 Hybrid Workflow Best Practices

The real power of ComfyUI lies in its ability to chain multiple AI models together into complex workflows. Here's what a typical Image-to-Video creation process looks like:

3.1 Hybrid Workflow Architecture Design

Workflow Diagram:

[文本提示词]
    ↓
[Nano Banana Pro 节点] → 生成高质量封面图像 (4K)
    ↓
[图像预处理节点] → 调整分辨率到 1280x720
    ↓
[Sora-2 i2v 节点] → 将图像转为 10 秒视频
    ↓
[视频后处理节点] → 添加转场和音效
    ↓
[最终输出] → MP4 视频文件

Node Configuration Example:

3.2 Actual Workflow Code Example

Minimal ComfyUI Workflow JSON (Core Nodes):

{
  "nodes": [
    {
      "id": 1,
      "type": "TextInput",
      "properties": {
        "text": "A cyberpunk city with neon lights and flying cars"
      }
    },
    {
      "id": 2,
      "type": "GoogleGeminiImage",
      "properties": {
        "api_key": "YOUR_API_KEY",
        "prompt": "@1.text",
        "resolution": "3840x2160"
      }
    },
    {
      "id": 3,
      "type": "ImageResize",
      "properties": {
        "input": "@2.output",
        "target_size": "1280x720"
      }
    },
    {
      "id": 4,
      "type": "OpenAISora2",
      "properties": {
        "api_key": "YOUR_OPENAI_KEY",
        "input_image": "@3.output",
        "duration": 10
      }
    }
  ]
}

{
  "version": "1.0",
  "workflow_name": "Nano_Banana_to_Sora2_Video",
  "nodes": [
    {
      "id": 1,
      "type": "TextInput",
      "properties": {
        "text": "A cyberpunk city with neon lights and flying cars",
        "negative_prompt": "blurry, low quality, distorted"
      }
    },
    {
      "id": 2,
      "type": "GoogleGeminiImage",
      "properties": {
        "api_key": "YOUR_API_KEY",
        "api_endpoint": "https://api.apiyi.com/v1/models/gemini-2.0-flash-exp:generateImage",
        "prompt": "@1.text",
        "negative_prompt": "@1.negative_prompt",
        "resolution": "3840x2160",
        "aspect_ratio": "16:9",
        "guidance_scale": 7.5,
        "num_inference_steps": 30
      }
    },
    {
      "id": 3,
      "type": "ImageResize",
      "properties": {
        "input": "@2.output",
        "target_size": "1280x720",
        "resize_mode": "crop_center",
        "interpolation": "lanczos"
      }
    },
    {
      "id": 4,
      "type": "OpenAISora2Pro",
      "properties": {
        "api_key": "YOUR_OPENAI_KEY",
        "input_image": "@3.output",
        "prompt": "@1.text",
        "duration": 10,
        "resolution": "1280x720",
        "motion_intensity": "medium",
        "camera_motion": "pan_right",
        "frame_rate": 30
      }
    },
    {
      "id": 5,
      "type": "VideoExport",
      "properties": {
        "input": "@4.output",
        "format": "mp4",
        "codec": "h264",
        "bitrate": "5000k",
        "output_path": "./outputs/final_video.mp4"
      }
    }
  ],
  "connections": [
    {"from": 1, "to": 2, "type": "text"},
    {"from": 2, "to": 3, "type": "image"},
    {"from": 3, "to": 4, "type": "image"},
    {"from": 4, "to": 5, "type": "video"}
  ]
}

🎯 Performance Optimization Tip: When handling batch tasks, it's a good idea to separate the Nano Banana Pro and Sora-2 API calls into different queues. By using the load balancing features of the APIYI (apiyi.com) platform, you can process image and video generation in parallel, boosting overall workflow efficiency by about 40%.

4. Frequently Asked Questions (FAQ)

Q1: Can Nano Banana Pro and Sora-2 be run entirely locally?

Nano Banana Pro: ❌ Local deployment isn't supported; it must be accessed via API calls.
Sora-2: ❌ Local deployment isn't supported; OpenAI hasn't open-sourced the model weights.
Alternatives: You can use open-source models like Stable Diffusion (for images) and Open-Sora (for video) to run things locally.

Q2: How can I reduce Sora-2 API call costs?

🎯 Cost Management Tip: For users spending over $500/month, we recommend ordering an enterprise package through the APIYI (apiyi.com) platform. The platform offers tiered pricing and prepaid discounts, while also supporting unified billing for both Nano Banana Pro and Sora-2, simplifying your financial process.

Q3: Is the video resolution generated by Sora-2 limited in ComfyUI?

Sora-2 Standard:

• ✅ Supported: 720×1280, 1280×720
• ❌ Not Supported: 1080p, 4K

Sora-2 Pro:

• ✅ Supported: 720×1280, 1280×720, 1024×1792, 1792×1024
• ❌ Not Supported: Custom resolutions

Solution: Use super-resolution nodes in ComfyUI (like ESRGAN) to upscale 720p video to 1080p or 4K.

Q4: How do I automatically pass images from Nano Banana Pro to Sora-2?

When setting up your ComfyUI workflow, keep these key points in mind:

# 错误示例:直接连接会导致分辨率不匹配
[Nano Banana Pro] → [Sora-2 i2v]  # ❌ 可能报错

# 正确示例:中间加入分辨率转换节点
[Nano Banana Pro] → [Image Resize] → [Sora-2 i2v]  # ✅ 推荐

Resolution Matching Table:

Q5: Which ComfyUI versions support these new models?

5. Summary and Recommendations

5.1 Quick Recap of Key Questions

Q: What's the simplest way to connect the Nano Banana Pro API to ComfyUI?
A: Just use the official Partner Nodes. It only takes three steps:

1. Update ComfyUI to the nightly version.
2. Search for and add the "Google Gemini Image" node.
3. Enter your API key and prompt, and you're all set.

Q: Can I use the Sora-2 video model in ComfyUI?
A: Yes, but you must use API nodes. Local model deployment isn't supported. Both official and third-party developers provide ready-to-use Sora-2 Custom Nodes.

5.2 Recommendations for Different Scenarios

🎯 Pro Tip: For users who need to use both Nano Banana Pro and Sora-2 simultaneously, we recommend using the APIYI (apiyi.com) platform for unified API management. It allows you to access both models with a single API Key, simplifying key management while providing a unified monitoring dashboard and cost analysis tools. Head over to apiyi.com to register—new users get $20 in free credits to test out their workflows.

5.3 Further Learning Resources

Official Documentation:

• ComfyUI Official Docs: docs.comfy.org/tutorials/partner-nodes
• Nano Banana Pro API Reference: ai.google.dev/gemini-api
• Sora-2 API Documentation: platform.openai.com/docs/guides/sora

Community Resources:

• ComfyUI Workflow Sharing: comfyui.org/workflows
• GitHub Custom Nodes Collection: github.com/ComfyUI-Workflow/awesome-comfyui

Video Tutorials:

• "Complete Guide to Nano Banana Pro in ComfyUI" – YouTube
• "Practical Demo: Sora-2 API Nodes" – Bilibili

Through the detailed explanations in this guide, you should now have a solid handle on how to integrate Nano Banana Pro and Sora-2 into ComfyUI. Whether you choose the simplicity of official nodes, the flexibility of custom nodes, or the enterprise-grade service of a third-party platform, you can pick the path that best fits your actual needs.

Now, fire up ComfyUI and start your AI creative journey!

Recommended Reading:

• Advanced ComfyUI: Building Automated Video Generation Workflows
• Nano Banana Pro vs. Midjourney: A Comparison of Enterprise Image Generation Models
• Sora-2 API Cost Optimization: How to Cut Video Generation Fees by 70%