ComfyUI-OllamaGemini

★ 171

多模型接入多模态（音视频）LLM集成ComfyUI扩展

为ComfyUI提供对Ollama、Gemini、OpenAI、Claude、Qwen的统一接入，支持视频与音频，简化多模型切换与多模态生成流程。

💡 在ComfyUI中一键调用多家模型进行音视频与文本生成。

🍴 29 Forks💻 Python🔄 2026-03-01

🔗 GitHub 原文

📦

网盘下载

复制链接后前往夸克网盘下载

https://pan.quark.cn/s/8f9eee5e2cdb

📦 requirements.txt

google-generativeai
Pillow
requests

📄 README

🤖 Gemini  ·  🧠 OpenAI  ·  🎭 Claude  ·  🦙 Ollama  ·  🌐 Qwen

_{Happy Creators}

_{Hours Saved Daily}

_{AI Providers}

_{Prompt Templates}

$\huge\textsf{\textcolor{ff6b35}{See It In Action}}$

https://github.com/user-attachments/assets/6ffba8bc-47e9-42c5-be98-5849ffb03547

🎞️ View More Examples

_{500+ Styles}

_{FLUX Resolutions}

_{SVG Conversion}

$\huge\textsf{\textcolor{c41e3a}{Why Creators Love Us}}$

$\textsf{\textcolor{c41e3a}{😫 Before}}$

- 5 different extensions to manage
- 5 different config files
- Inconsistent prompt formats
- Hours wasted switching tools
- Frequent compatibility issues

→

$\textsf{\textcolor{7ed321}{✨ After}}$

+ ONE unified extension
+ ONE config for all APIs
+ Smart prompt optimization
+ Instant provider switching
+ Always up-to-date

$\huge\textsf{\textcolor{ffd700}{Powerful Features}}$

Gemini

_{2.0 Pro • Flash • 1.5}

ChatGPT

_{GPT-4o • 4-Turbo • 3.5}

Claude

_{3.7 • 3.5 Sonnet • Opus}

Ollama

_{Any Local Model}

Qwen

_{Max • Plus • Turbo}

Veo 3.1 Video

_{Text/Image to Video + Extend}

Background Removal

_{BRIA RMBG hair-level detail}

Imagen 4

_{Google’s latest image model}

Gemini Banana Pro

_{Advanced image editing}

FLUX Resolutions

_{Perfect sizing for every model}

500+ Art Styles

_{🎨 Curated artistic presets}

Smart Prompts

_{AI-enhanced engineering}

Multi Prompt

_{Batch processing workflows}

_{Latest AI for Object Detection & Segmentation}

SAM3 Direct Text

_{Prompts: “sun”, “face”, “cat”}

Auto-Install UI

_{Popups & Browser Auto-Open}

BiRefNet Matting

_{Hair-level edge quality}

Precision Controls

_{Confidence & Feathering}

$\huge\textsf{\textcolor{ffd700}{500+ Curated Art Styles}}$

_Cinema

_{80+ styles}

_{Fine Art}

_{120+ styles}

_Gaming

_{60+ styles}

_Photo

_{90+ styles}

_Fantasy

_{100+ styles}

🔥 View Popular Style Categories

| Category | Styles | Examples |

|:———|:——:|:———|

| 🎬 Cinematic | 80+ | Film Noir, Blade Runner, Spielberg, Nolan, Wes Anderson |

| 🖼️ Fine Art | 120+ | Van Gogh, Monet, Picasso, Rembrandt, Caravaggio |

| 🎮 Digital Art | 60+ | Cyberpunk, Synthwave, Vaporwave, Pixel Art, 3D Render |

| 📸 Photography | 90+ | Portrait, Landscape, Street, Fashion, Product |

| ✨ Fantasy | 100+ | Epic Fantasy, Dark Fantasy, Fairy Tale, Mythological |

| 🎌 Anime | 50+ | Studio Ghibli, Makoto Shinkai, Trigger, Mappa |

$\huge\textsf{\textcolor{f5a623}{Quick Start}}$

⚡ Install in under 30 seconds

$\textsf{\textcolor{7ed321}{▶ Recommended}}$

ComfyUI Manager *(One-Click)*

1. Open ComfyUI Manager
2. Search "OllamaGemini"  
3. Click Install ✓
4. Restart ComfyUI

$\textsf{\textcolor{ff6b35}{▷ Manual}}$

Git Clone

cd ComfyUI/custom_nodes
git clone https://github.com/al-swaiti/ComfyUI-OllamaGemini.git
pip install -r requirements.txt

🔑 API Configuration

{
  "GEMINI_API_KEY": "your_key",      // 🆓 aistudio.google.com
  "OPENAI_API_KEY": "your_key",      // 💰 platform.openai.com
  "ANTHROPIC_API_KEY": "your_key",   // ⚠️ console.anthropic.com
  "OLLAMA_URL": "http://localhost:11434",  // 🆓 Local
  "QWEN_API_KEY": "your_key"         // ⚠️ dashscope.console.aliyun.com
}

$\huge\textsf{\textcolor{c41e3a}{20+ Prompt Templates}}$

_{Extensively researched • Model-optimized • Professional results}

🎬 Video Generation

| Template | Description |

|:———|:————|

| Veo3-TextToVideo | Google Veo 3.1 with composition, camera, subject, action & native audio |

| Veo3-ReferenceImages | Reference image video preserving subject appearance |

| Veo3-Interpolation | First-to-last frame interpolation with motion paths |

| VideoGen | Professional cinematography: subject, action, lighting, style |

⚡ FLUX Models

| Template | Description |

|:———|:————|

| FLUX.1-dev | Hyper-detailed cinematographic with lighting & camera specs |

| FLUX.2-dev | Natural language following official BFL guide |

| FLUX.2-dev-Edit | Multi-reference editing for up to 10 images |

| FLUX.2-dev-JSON | Structured JSON for complex scenes |

| FLUXKontext | Context-aware editing with character consistency |

🎨 Image Generation

| Template | Description |

|:———|:————|

| SDXL | Premium comma-separated tags with artistic medium |

| Imagen4 | Structured, layered prompts for Google Imagen 4 |

| Z-Image-Turbo | 6B diffusion transformer for concept fusion |

| Qwen-Image-2512 | Photorealistic eliminating “AI look” |

| Upscale | Sharpness-maximizing enhancement |

🍌 Gemini Nano Banana Pro

| Template | Description |

|:———|:————|

| GeminiNanaBananaEdit | Mask-free contextual editing |

| NanaBananaPro | Gemini 3 Pro Image with narrative style |

| NanaBananaPro-Edit | Advanced editing with multi-image composition |

| NanaBananaPro-Pro | Professional 4K asset production |

$\huge\textsf{\textcolor{c41e3a}{Support This Project}}$

⏰ 500+ hours of development

💝 Your support keeps it FREE for everyone

⭐ Every star & donation means the world 💖

Abdallah Al-Swaiti

_{🇯🇴 Amman, Jordan}

*”I built this because I was frustrated switching between 5 different AI tools.*

*Now, 150+ creators use it daily. If this helps your workflow, consider supporting!”*

$\large\textsf{\textcolor{f5a623}{Connect}}$

FREE • Open Source • MIT License

Made with ❤️ in Jordan 🇯🇴