feat: Add ModelsLab provider for image and video generation

[adhikjoshi]

Summary

Adds ModelsLab as a new provider for both image and video generation in OpenMontage.

What's Included

ModelsLab Image Generation (tools/graphics/modelslab_image.py)

• Flux models (flux-v1-schnell, flux-v1-dev, flux-pro, flux-sdxl)
• SDXL and other models (sdxl, turbo-v2, realistic-v5, anime-v6)
• Image-to-image generation with reference image URL
• Negative prompts, custom size, seed control
• Async polling for generation status

ModelsLab Video Generation (tools/video/modelslab_video.py)

• wan2.1 video generation model
• Text-to-video and image-to-video operations
• Async polling for video completion

Usage

# Add to .env
MODELSLAB_API_KEY=your-key-here

# Then in your agent:
"Generate an image of a sunset over mountains using ModelsLab"

Provider Comparison

Provider	Image Cost	Video Cost	Best For
fal.ai (FLUX)	~$0.03-0.05	~$0.05/sec	Broad coverage
ModelsLab	~$0.02-0.05	~$0.05/sec	Cost-effective alternative
OpenAI (DALL-E 3)	~$0.04-0.08	N/A	Quality

Pricing

ModelsLab offers competitive pricing on Flux and SDXL models. See modelslab.com for current rates.

Testing Checklist

• python -c "from tools.tool_registry import registry; registry.discover(); print(registry.get('modelslab_image').get_status())"
• python -c "from tools.tool_registry import registry; registry.discover(); print(registry.get('modelslab_video').get_status())"
• Image generation with different models
• Video generation (text-to-video)
• Async polling behavior

Notes

• Follows existing provider patterns (flux_image.py, minimax_video.py)
• Env var: MODELSLAB_API_KEY
• No new dependencies required
• Tools auto-discovered via tool_registry.py