A FoundryVTT module that integrates Runware AI image generation directly into actor sheets. Generate high-quality AI images for your NPCs and player characters with ease!
- 🎨 Easy Access: Generate images directly from actor sheets with a button in the header bar (next to the prototype token button)
- 🤖 Multiple AI Models: Support for various AI models including Stable Diffusion, SDXL, and custom models from CivitAI
- 🧰 GM Presets: Game Masters can curate shared presets (model, LoRA, VAE, embeddings) for players to apply instantly
- 🎯 Advanced Controls:
- Positive and negative prompts
- Adjustable image dimensions (256x256 to 2048x2048)
- LoRA model support for style adaptation
- CFG Scale, inference steps, and seed control
- Generate multiple images at once (1-4) with an in-app gallery to pick your favorite result
- 💾 Organized Storage: Portraits and token variants are automatically saved to
modules/runware-imagegen/images/actor-name/(tokens live in the/tokenssubfolder) - 🖼️ Quick Application: Instantly apply the chosen render as the actor portrait, with multi-image preview selection when multiple results are generated
- 🪄 Token Ready: Backgrounds are removed automatically via Runware RMBG v2.0 and saved as autorenamed token images alongside the portrait
- 🔐 Secure: API keys stored securely in world settings
- In FoundryVTT, go to Add-on Modules
- Click Install Module
- Search for "Runware AI Image Generator"
- Click Install
- Download the latest release from the [GitHub repository]
- Extract the zip file into your FoundryVTT
Data/modulesdirectory - Restart FoundryVTT
- Enable the module in your world's Module Management settings
- Visit Runware AI
- Sign up for an account
- Navigate to your API settings
- Generate an API key
- In FoundryVTT, go to Settings → Configure Settings → Module Settings
- Find Runware AI Image Generator
- Enter your API key in the Runware API Key field
- Optionally configure default settings:
- Default Model: The AI model to use by default (e.g.,
runware:100@1) - Image Width: Default width for generated images (512px recommended)
- Image Height: Default height for generated images (512px recommended)
- Number of Results: How many images to generate per request (1-4)
- Default Model: The AI model to use by default (e.g.,
- Open an Actor Sheet for any NPC or Character
- Click the "Generate Image" button (🎨 palette icon) in the top bar next to the prototype token button
- Enter a Prompt: Describe the image you want to generate
- Example: "A wise elderly wizard with a long white beard, wearing blue robes, fantasy art style"
- Select a Model: Choose from the suggested models or enter a custom model ID
- Click "Generate Image"
- Wait for Generation: The module will display a loading indicator
- Review the Results: If multiple images were requested, a gallery lets you preview and choose your favorite; the selected image is then shown in a confirmation dialog
- Confirm or Save: The chosen portrait is stored in
modules/runware-image-generator/images/[actor-name]/and a matching background-free token is generated automatically inmodules/runware-image-generator/images/[actor-name]/tokens/
Click Show Advanced Options in the dialog to access:
Add style-specific fine-tuning to your images:
- LoRA Model: Enter a LoRA model ID (e.g.,
civitai:12345@67890) - Weight: Adjust the influence of the LoRA (0.0 - 2.0, default 1.0)
- Inference Steps: More steps = higher quality but slower (20-50 recommended)
- CFG Scale: How closely to follow the prompt (7.0 recommended)
- Seed: For reproducible results, enter a specific number
- Adjust width and height independently (256-2048px)
- Use multiples of 64 for best results
- Square images (512x512) typically work best
The module includes quick-select buttons for popular models:
- Stable Diffusion 1.5 (
runware:100@1) - Stable Diffusion XL (
runware:101@1) - Realistic Vision (
civitai:4201@130072) - DreamShaper (
civitai:4384@128713)
You can use any model from:
- Runware Models: Format
runware:MODEL_ID@VERSION - CivitAI Models: Format
civitai:MODEL_ID@VERSION_ID
Visit CivitAI to browse thousands of community models.
Generated images are saved in an organized directory structure:
Data/
modules/
runware-image-generator/
images/
warrior_character/
image_1.png
image_2.png
image_3.png
tokens/
token_1.png
npc_shopkeeper/
image_1.png
image_2.png
tokens/
token_1.png
- Actor names are sanitized (special characters replaced with underscores)
- Images are numbered sequentially
- Images persist across sessions
- You can access these files directly via the FilePicker
- Background-removed token images are stored alongside portraits under the
tokens/subdirectory
- Be specific about appearance, style, and mood
- Include details like clothing, accessories, lighting
- Mention art style (e.g., "fantasy art", "realistic", "anime style")
- Use descriptive adjectives
Example Good Prompt:
A fierce female warrior with red hair in a ponytail, wearing silver plate armor
with gold trim, holding a flaming sword, dramatic lighting, fantasy art style,
detailed face, heroic pose
Use negative prompts to avoid unwanted elements:
- Common exclusions:
blurry, low quality, deformed, ugly, text, watermark - Style exclusions:
cartoon, anime(if you want realistic) - Content exclusions: Specific unwanted objects or features
You can emphasize parts of your prompt:
- Use parentheses:
(important detail:1.2)increases weight - Use brackets:
[less important detail:0.8]decreases weight
Solution: Go to Module Settings and enter your valid Runware API key.
Possible causes:
- Invalid API key
- Insufficient credits in your Runware account
- Invalid model ID
- Network connection issues
Solutions:
- Verify your API key is correct
- Check your Runware account balance
- Try a different model from the suggestions
- Check browser console for detailed error messages
Possible causes:
- Insufficient permissions on the Data directory
- File path issues
Solutions:
- Ensure FoundryVTT has write permissions to the Data folder
- Check browser console for specific errors
- Try generating with a different actor name
Possible causes:
- Module not enabled
- Not viewing as actor owner
- Sheet template incompatibility
Solutions:
- Verify module is enabled in Module Management
- Check that you have OWNER permission on the actor
- Try with a different actor sheet type
This module uses the Runware SDK to communicate with the Runware API. The SDK is loaded dynamically via CDN.
- Text-to-Image Generation: Primary image generation
- Base64 Output: Images are received as base64 for local saving
- Multiple Models: Support for Runware and CivitAI models
- LoRA Support: Style adaptation via LoRA models
- Advanced Parameters: CFG Scale, steps, seed control
- Background Removal: Automated RMBG v2.0 workflow to create transparent token assets
- Runware AI: https://runware.ai - AI image generation API
- Runware SDK: https://github.com/runware/sdk-js
- FoundryVTT: https://foundryvtt.com
This module is released under the MIT License. See LICENSE file for details.
For issues, feature requests, or questions:
- GitHub Issues: [Repository Issues Page]
- FoundryVTT Discord: Look for the module support channel
- Added multi-result preview gallery so users can pick the best generation before saving
- Automatic background removal via Runware RMBG v2.0 with token images saved alongside portraits
- Enlarged image preview dialog for easier inspection of generated art
- Introduced GM-managed presets covering model, LoRA, VAE, and embeddings shared with players
- Initial release
- Basic text-to-image generation
- Actor sheet integration
- Multiple model support
- LoRA support
- Advanced parameter controls
- Automatic image saving and organization
- Quick portrait application
Future planned features:
- Image-to-image generation
- Batch generation for multiple actors
- Image history browser
- Preset prompt templates
- Style library
Enjoy creating amazing character art with AI! 🎨✨