ComfyUI-Google-AI-Studio

ComfyUI Google AI Studio Nodes

This custom node package provides integration with Google AI Studio's latest APIs for ComfyUI, including Text-to-Speech (TTS), Text Generation, and Image Generation (Nano Banana) using the google-genai Google Gen AI SDK (Google AI Studio).

Features

• Text Generation: Generate creative text, stories, essays using Gemini models
• Image Generation: Create stunning images using Google's Gemini (free) and Imagen (paid) models
• Image Editing: Edit existing images using Gemini 2.5 Flash with multimodal input support
• Batch Processing: Process multiple images simultaneously in a single API call
• Google AI Studio TTS: Convert text to speech using Google's Gemini TTS models
• Multi-Speaker TTS: Generate conversations with multiple distinct voices
• 30+ Voice Options: Choose from a wide variety of high-quality voices with style descriptions

Installation

1. Clone this repository into your ComfyUI custom_nodes directory:

   cd ComfyUI/custom_nodes
   git clone https://github.com/BuffMcBigHuge/ComfyUI-Google-AI-Studio.git
   

2. Install the required dependencies:

   cd ComfyUI-Google-AI-Studio
   pip install -r requirements.txt
   

3. Get your Google AI Studio API key:

   • Visit https://aistudio.google.com/
   • Sign up or log in
   • Generate an API key
   • Keep this key secure

4. Restart ComfyUI

Nodes

Google AI Studio Text Generator

Generates text using Google's Gemini models for creative writing, essays, code, and more.

Inputs:

• prompt: The text prompt to generate from (multiline supported)
• api_key: Your Google AI Studio API key
• model: Choose from Gemini models (gemini-2.5-flash, gemini-2.5-pro, gemini-3.1-pro-preview, etc.)
• system_instruction (optional): Guide the model's behavior and tone
• temperature (optional): Controls creativity (0.0-2.0, default 0.7)
• max_output_tokens (optional): Maximum tokens to generate (1-8192, default 1024)
• thinking_level (optional): Reasoning depth for Gemini 2.5/3 (off, low, medium, high)

Output:

• text: Generated text content

Google AI Studio Image Generator

Creates images using Google's Gemini and Imagen models from text descriptions.

Available Models:

• gemini-3.1-flash-image-preview - Free tier (default, Nano Banana 2)
• gemini-3-pro-image-preview - Free tier (Nano Banana Pro)
• gemini-2.5-flash-image - Free tier (Nano Banana, efficient)
• imagen-4.0-fast-generate-001 - Paid tier (Imagen 4 Fast)
• imagen-4.0-generate-001 - Paid tier (Imagen 4)
• imagen-4.0-ultra-generate-001 - Paid tier (Imagen 4 Ultra)
• imagen-3.0-fast-generate-001 - Paid tier (Imagen 3 Fast)

Inputs:

• prompt: Description of the image to generate or edit
• api_key: Your Google AI Studio API key
• model: Choose from available models above
• input_image (optional): Input image(s) for editing/modification - supports batch processing (⚠️ Gemini models only)
• negative_prompt (optional): What you don't want in the image (works with all models)
• aspect_ratio (optional): Image proportions (⚠️ Imagen only - ignored for Gemini)
• safety_filter_level (optional): Content filtering (⚠️ Imagen only - ignored for Gemini)
• person_generation (optional): People generation controls (⚠️ Imagen only - ignored for Gemini)

Google AI Studio TTS

Converts text to speech using Google's Gemini TTS models.

Inputs:

• text: The text to convert to speech (multiline supported)
• api_key: Your Google AI Studio API key
• model: Choose between gemini-2.5-flash-preview-tts or gemini-2.5-pro-preview-tts
• voice: Select from 30+ available voices (Kore, Puck, Zephyr, etc.)
• instruction_prefix (optional): Add style instructions like "Say cheerfully:" or "Read slowly:"

Output:

• audio: Audio data compatible with ComfyUI audio nodes

Google AI Studio Multi-Speaker TTS

Generates conversations with multiple distinct voices.

Inputs:

• transcript: Multi-line text with speaker names (format: "Speaker Name: dialogue")
• api_key: Your Google AI Studio API key
• model: TTS model selection
• speaker1_name / speaker1_voice: First speaker configuration
• speaker2_name / speaker2_voice: Second speaker configuration
• speaker3_name / speaker3_voice (optional): Third speaker
• speaker4_name / speaker4_voice (optional): Fourth speaker

Output:

• audio: Multi-speaker audio data

Model Capabilities:

| Feature | Gemini Models (Free) | Imagen Models (Paid) | |---------|---------------------|---------------------| | Text-to-image | ✅ | ✅ | | Image-to-image editing | ✅ | ❌ | | Negative prompts | ✅ (in main prompt) | ✅ (separate parameter) | | Aspect ratio control | ❌ | ✅ | | Safety filter levels | ❌ (built-in safety) | ✅ | | Person generation control | ❌ | ✅ |

Output:

• image: Generated image compatible with ComfyUI image nodes

Available Voices

The following 30 voices are available, each with distinct characteristics:

| Voice | Style | Voice | Style | Voice | Style | |-------|--------|-------|--------|-------|--------| | Zephyr | Bright | Puck | Upbeat | Charon | Informative | | Kore | Firm | Fenrir | Excitable | Leda | Youthful | | Orus | Firm | Aoede | Breezy | Callirrhoe | Easy-going | | Autonoe | Bright | Enceladus | Breathy | Iapetus | Clear | | Umbriel | Easy-going | Algieba | Smooth | Despina | Smooth | | Erinome | Clear | Algenib | Gravelly | Rasalgethi | Informative | | Laomedeia | Upbeat | Achernar | Soft | Alnilam | Firm | | Schedar | Even | Gacrux | Mature | Pulcherrima | Forward | | Achird | Friendly | Zubenelgenubi | Casual | Vindemiatrix | Gentle | | Sadachbia | Lively | Sadaltager | Knowledgeable | Sulafat | Warm |

Supported Languages

The TTS models automatically detect input language and support 24 languages including:

• English (US, India)
• Spanish (US)
• French (France)
• German (Germany)
• Italian (Italy)
• Portuguese (Brazil)
• Japanese (Japan)
• Korean (Korea)
• Chinese (Simplified, Traditional)
• Hindi (India)
• Arabic (Egyptian)
• Russian (Russia)
• And more...

Usage Examples

Text Generation Example

1. Add a "Google AI Studio Text Generator" node
2. Enter your prompt: "Write a creative short story about a robot learning to paint"
3. Enter your API key
4. Select model: gemini-2.5-flash
5. Adjust temperature for creativity (0.7 for balanced, 1.5 for very creative)
6. Connect output to text display or save node

Image Generation Example

1. Add a "Google AI Studio Image Generator" node
2. Enter your prompt: "A serene Japanese garden with cherry blossoms and a small pond"
3. Enter your API key
4. Select model: gemini-3.1-flash-image-preview (free tier, default)
5. For advanced controls (paid tier), switch to imagen-4.0-generate-001 and adjust aspect ratio, safety settings
6. Connect output to image preview or save node

Image Editing Example

1. Add a "Google AI Studio Image Generator" node
2. Connect an input image to the input_image parameter
3. Enter your editing prompt: "Add a rainbow in the sky and make it look more vibrant"
4. Enter your API key
5. Select a Gemini model (2.5 Flash or 3.1 Flash Image - image editing only works with Gemini models)
6. Connect output to image preview or save node

Batch Image Processing Example

1. Add a "Google AI Studio Image Generator" node
2. Connect a batch of images to the input_image parameter (e.g., from an image batch loader)
3. Enter your editing prompt: "Apply vintage film effect to these images"
4. Enter your API key
5. Select a Gemini model - all images in the batch will be sent in a single API call
6. Connect output to image preview or save node

Basic TTS Example

1. Add a "Google AI Studio TTS" node
2. Enter your text: "Welcome to my podcast! Today we're discussing AI."
3. Enter your API key
4. Select voice: Kore
5. Connect output to audio preview or save node

Multi-Speaker Conversation Example

1. Add a "Google AI Studio Multi-Speaker TTS" node
2. Enter transcript:

   Host: Welcome to Tech Talk!
   Guest: Thanks for having me on the show.
   Host: Let's dive into today's topic.
   

3. Configure speakers:
   • Speaker 1: Host with voice Kore
   • Speaker 2: Guest with voice Puck
4. Connect output to audio nodes

Style Control Example

Use the instruction_prefix to control speech style:

• "Say cheerfully:" - Upbeat delivery
• "Read slowly and clearly:" - Careful pronunciation
• "Speak in a whisper:" - Quiet, intimate tone
• "Say with excitement:" - Energetic delivery

Technical Details

• Audio Format: 24kHz, 16-bit PCM, mono/stereo
• Context Window: 32k tokens maximum
• Output Format: ComfyUI-compatible audio tensors
• API Rate Limits: Subject to Google AI Studio quotas

Troubleshooting

Common Issues

1. "Google AI SDK not installed"

   • Run: pip install google-genai

2. "API key is required"

   • Get your key from https://aistudio.google.com/
   • Ensure the key is entered correctly

3. "TTS generation failed"

   • Check your internet connection
   • Verify API key is valid and has quota remaining
   • Ensure text is within the 32k token limit

4. Audio playback issues

   • Ensure ComfyUI audio nodes are properly connected
   • Check that audio preview nodes support the format

API Limits

• Free tier: Limited requests per minute
• Check your usage at Google AI Studio

Support

For issues specific to these nodes:

• Check the Issues page
• Review the troubleshooting section above

For Google AI Studio API issues:

• Visit Google AI Studio documentation
• Check the Google AI Forum

License

This project is licensed under the MIT License - see the LICENSE file for details.

Changelog

v1.2.0 - Model Updates & Thinking Control

🆕 New Features

• Thinking Level: Optional thinking_level parameter for Text Generator (off, low, medium, high) - controls reasoning depth for Gemini 2.5/3 models

🔄 Model Updates (per Gemini API changelog & deprecations)

• Text Generator: Added gemini-3.1-pro-preview, gemini-3.1-flash-lite-preview, gemini-3-flash-preview, gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite. Removed deprecated gemini-3-pro-preview, gemini-2.0-flash-001. Default: gemini-2.5-flash
• Image Generator: Added gemini-3.1-flash-image-preview (Nano Banana 2), imagen-4.0-fast-generate-001, imagen-4.0-ultra-generate-001. Default: gemini-3.1-flash-image-preview

---

v1.1.0 - Image Editing & Batch Processing

New Feature: Image editing with Gemini 2.5 Flash - example shows transforming Darth Vader with the prompt "put the character on the beach"

🆕 New Features

• Image Editing Support: Added Gemini 2.5 Flash image editing capabilities

  • Edit existing images with text prompts using multimodal input
  • Supports single images or batch processing
  • Works with all ComfyUI image nodes and workflows

• Gemini 3 Models: Added support for gemini-3-pro-preview and gemini-3-pro-image-preview.

• Gemini 2.5 Flash Image Model: Updated to gemini-2.5-flash-image.

• Imagen 4: Added imagen-4.0-generate-001.

  • Latest Google image generation model
  • Free tier access with advanced capabilities
  • Enhanced image quality and prompt understanding

🔄 Improvements

• Enhanced Image Generator node with optional input_image parameter
• Updated API integration with proper multimodal Content/Part structure
• Improved batch processing efficiency for multiple images
• Better error handling for image conversion and processing

---

Note: This is an unofficial community-created integration. Google AI Studio is a product of Google LLC.