ComfyUI Extension: ComfyUI-Gemini-Prompt-Studio
One node generates both T2I & I2V positive/negative conditioning in a single click.
Custom Nodes (0)
README
ComfyUI-Gemini-Prompt-Studio
Enhance your ComfyUI prompts with Google Gemini's dynamic “artist persona” system.
One node generates both T2I & I2V positive/negative conditioning in a single click.
<!-- 中文 -->
ComfyUI Gemini 提示词工作室节点
借助 Google Gemini 的「动态艺术人格」系统,一键生成文生图 + 图生视频的正反提示词与条件。
<!-- English -->
Features
- 5 built-in artist personas (film director, art historian, concept artist, novelist, creative director)
- Automatic quality modifiers & negative prompts
- Seed-controlled randomness for reproducible results
- Outputs 4 × conditioning + 4 × raw text + status string
<!-- 中文 -->
主要功能
- 内置 5 种艺术人格(电影导演、艺术史学家、概念设计师、文学家、广告创意总监)
- 自动附加画质增强与负面提示词
- 支持种子复现,结果可复现
- 同时输出 4 组条件 + 4 段文本 + 状态信息
<!-- English -->
Installation
✅ Recommended: ComfyUI-Manager
- Open Manager → Install Custom Nodes
- Search
Gemini Prompt Studio
- Click Install
🖥️ Manual
cd ComfyUI/custom_nodes
git clone https://github.com/DaLongZhuaZi/ComfyUI-Gemini-Prompt-Studio.git
cd ComfyUI-Gemini-Prompt-Studio
pip install -r requirements.txt
Restart ComfyUI.
<!-- 中文 -->
安装方式
✅ 推荐:ComfyUI-Manager
- 打开 管理器 → 安装自定义节点
- 搜索
Gemini Prompt Studio
- 点击 安装
🖥️ 手动安装
cd ComfyUI/custom_nodes
git clone https://github.com/DaLongZhuaZi/ComfyUI-Gemini-Prompt-Studio.git
cd ComfyUI-Gemini-Prompt-Studio
pip install -r requirements.txt
重启 ComfyUI 即可。
<!-- English -->
Quick Start
Step 1: Get API Key
Get a Google AI Studio API key (free tier available)
Step 2: Add Node to Workflow
- Drag node
Gemini Prompt Studio (T2I+I2V) ✨
into your ComfyUI workflow - The node appears in category: Magic Tools/Generation
Step 3: Configure Inputs
| Input | Type | Description | Example |
|-------|------|-------------|---------|
| clip | CLIP | Connect from your model's CLIP encoder | Required connection |
| api_key | STRING | Your Google AI Studio API key | xxxxxxx...
|
| model_name | DROPDOWN | Gemini model selection | gemini-2.5-flash
(recommended) |
| keywords | STRING | Core subject/scene description | A beautiful girl in a white dress, sitting by a window
|
| animation_effect | STRING | How the image should animate (optional) | She slowly turns her head to look at the camera, wind blows her hair
|
| seed | INT | Random seed for reproducible results | 42
|
Step 4: Connect Outputs
The node provides 9 outputs:
| Output | Type | Description | Connect To | |--------|------|-------------|------------| | T2I_POSITIVE | CONDITIONING | Enhanced positive prompt for text-to-image | KSampler → positive | | T2I_NEGATIVE | CONDITIONING | Enhanced negative prompt for text-to-image | KSampler → negative | | I2V_POSITIVE | CONDITIONING | Enhanced positive prompt for image-to-video | AnimateDiff → positive | | I2V_NEGATIVE | CONDITIONING | Enhanced negative prompt for image-to-video | AnimateDiff → negative | | t2i_pos_text | STRING | Raw positive text (for preview/debug) | Text Display Node | | t2i_neg_text | STRING | Raw negative text (for preview/debug) | Text Display Node | | i2v_pos_text | STRING | Raw I2V positive text (for preview/debug) | Text Display Node | | i2v_neg_text | STRING | Raw I2V negative text (for preview/debug) | Text Display Node | | gemini_status | STRING | API call status and error messages | Text Display Node |
Step 5: Basic Workflow Connection
[Load Checkpoint] → CLIP → [Gemini Prompt Studio] → T2I_POSITIVE → [KSampler]
→ T2I_NEGATIVE → [KSampler]
Step 6: Advanced I2V Workflow
[Image] → [Gemini Prompt Studio] → I2V_POSITIVE → [AnimateDiff/SVD]
→ I2V_NEGATIVE → [AnimateDiff/SVD]
<!-- 中文 -->
快速上手
步骤 1:获取 API 密钥
申请 Google AI Studio API 密钥(有免费额度)
步骤 2:添加节点到工作流
- 拖拽节点
Gemini Prompt Studio (T2I+I2V) ✨
到你的 ComfyUI 工作流中 - 节点位于分类:Magic Tools/Generation
步骤 3:配置输入参数
| 输入参数 | 类型 | 说明 | 示例 |
|----------|------|------|------|
| clip | CLIP | 连接模型的 CLIP 编码器 | 必须连接 |
| api_key | STRING | 你的 Google AI Studio API 密钥 | xxxxxxx...
|
| model_name | 下拉选择 | Gemini 模型选择 | gemini-2.5-flash
(推荐) |
| keywords | STRING | 核心主题/场景描述 | 一个穿白裙的美丽女孩,坐在窗边
|
| animation_effect | STRING | 画面如何动起来(可选) | 她慢慢转头看向镜头,微风轻抚头发
|
| seed | INT | 随机种子,用于结果复现 | 42
|
步骤 4:连接输出
节点提供 9 个输出:
| 输出 | 类型 | 说明 | 连接到 | |------|------|------|--------| | T2I_POSITIVE | CONDITIONING | 增强的文生图正面提示词 | KSampler → positive | | T2I_NEGATIVE | CONDITIONING | 增强的文生图负面提示词 | KSampler → negative | | I2V_POSITIVE | CONDITIONING | 增强的图生视频正面提示词 | AnimateDiff → positive | | I2V_NEGATIVE | CONDITIONING | 增强的图生视频负面提示词 | AnimateDiff → negative | | t2i_pos_text | STRING | 原始正面文本(预览/调试用) | 文本显示节点 | | t2i_neg_text | STRING | 原始负面文本(预览/调试用) | 文本显示节点 | | i2v_pos_text | STRING | 原始 I2V 正面文本(预览/调试用) | 文本显示节点 | | i2v_neg_text | STRING | 原始 I2V 负面文本(预览/调试用) | 文本显示节点 | | gemini_status | STRING | API 调用状态和错误信息 | 文本显示节点 |
步骤 5:基础工作流连接
[加载检查点] → CLIP → [Gemini Prompt Studio] → T2I_POSITIVE → [KSampler]
→ T2I_NEGATIVE → [KSampler]
步骤 6:高级图生视频工作流
[图像] → [Gemini Prompt Studio] → I2V_POSITIVE → [AnimateDiff/SVD]
→ I2V_NEGATIVE → [AnimateDiff/SVD]
<!-- English -->
Example Workflow
Download examples/T2I+I2V.json
and examples/TexttoImage.json
→ drag into ComfyUI → Run.
<!-- 中文 -->
示例工作流
下载 examples/T2I+I2V.json
和 examples/TexttoImage.json
→ 拖进 ComfyUI → 运行。
<!-- English -->
Troubleshooting
| Problem | Solution |
|---------|----------|
| Node not visible | Make sure __init__.py
is in the repo root |
| ModuleNotFoundError: google.generativeai
| Re-run pip install -r requirements.txt
|
| API returns empty | Check key quota & safety filters; lower temperature or tweak prompt |
| Other API issues | Visit |
<!-- 中文 -->
常见问题
| 问题 | 解决方法 |
|------|----------|
| 节点未显示 | 确认 __init__.py
位于仓库根目录 |
| 提示缺少 google.generativeai
| 重新执行 pip install -r requirements.txt
|
| API 返回为空 | 检查密钥额度及安全过滤,可调低 temperature 或修改提示词 |
| 其他 API 相关问题 | 访问 Google AI Studio 文档排查 |
<!-- English -->
Contributing
Pull requests are welcome!
Please run python -m ruff check .
before submitting.
<!-- 中文 -->
参与贡献
欢迎提交 PR!
提 PR 前请执行 python -m ruff check .
保证代码风格一致。
<!-- English -->
License
MIT © 2025 DaLongzhuazi
<!-- 中文 -->
开源协议
MIT 许可证 © 2025 DaLongZhuaZi