feat: add image text model provider separation and fal.ai integration

[yoniebans]

Relates to:

• Optional image model provider to character and runtime #647
• fal.ai image generation #648

Risks

Low - Adds optional fal.ai integration for image generation. Existing functionality remains unchanged when new optional imageModelProvider is not set.

Background

What does this PR do?

• Adds fal.ai integration for image generation
• Introduces separate optional image model provider configuration
• Adds FAL_API_KEY and FAL_AI_LORA_PATH environment variables
• Implements fal.ai client setup and image generation logic

What kind of change is this?

Features (non-breaking change which adds functionality)

Documentation changes needed?

My changes require a change to the project documentation to document:

• New environment variables
• How to configure fal.ai image generation
• Using separate model providers for text and image generation

Testing

Where should a reviewer start?

1. agent/src/index.ts - Updated to include image generation plugin should one of the valid image generation model api keys be included.
2. packages/core/src/generation.ts - Review fal.ai integration. Changed to look at model provider on runtime instance. if modelProvider = imageModelProvider, use token otherwise use one of the image env vars set.
3. packages/core/src/models.ts - Added new entry for FAL model
4. packages/core/src/types.ts - Check new imageModelProvider type additions
5. packages/core/src/runtime.ts - Review imageModelProvider implementation. Updated to set imageModelProvider if one is provided in character config otherwise default to modelProvider.

Detailed testing steps

fal ai api testing:

1. Add FAL_API_KEY to .env
2. Configure character with imageModelProvider: "falai"
3. Test image generation with default settings
4. Test with custom FAL_AI_LORA_PATH
5. Verify existing providers still work when "falai" is not used
6. Test switching between different image providers

same modelProvider to imageModelProvider tests:

1. Set OPENAI_API_KEY
2. Set modelProvider to "openai" in character config
3. Both text generation and image generation goes through openai

different modelProvider to imageModelProvider tests:

1. Set OPENAI_API_KEY & FAL_API_KEY
2. Set modelProvider to "openai" and imageModelProvider to "falai" in character config
3. text generation goes through openai and image generation goes through falai

different modelProvider to imageModelProvider tests:

1. Set OPENAI_API_KEY & HEURIST_API_KEY
2. Set modelProvider to "openai" and imageModelProvider to "heurist" in character config
3. text generation goes through openai and image generation goes through heurist

Discord username

yoniebans