Model Swap Agent Added

[kailashahirwar]

Description

This PR adds a new Model Swap Agent that intelligently replaces models/people in images while preserving outfits and styling. The agent uses LangChain to extract person attributes from natural language prompts and generates professional model-swapped images using multiple AI models (Nano Banana, Nano Banana Pro, FLUX 2 Pro, FLUX 2 Flex).

Related Issue

Fixes #86

Type of Change

• New feature (non-breaking change that adds functionality)
• Documentation update

Component Affected

• AI Agents
• Image Generation (Nano Banana / FLUX.2)
• CLI Tools
• Documentation

Changes Made

• Added Model Swap Agent (tryon/agents/model_swap/):

  • agent.py: Main ModelSwapAgent class with LangChain integration
  • tools.py: LangChain tools for all supported models (Nano Banana, Nano Banana Pro, FLUX 2 Pro, FLUX 2 Flex)
  • __init__.py: Module exports

• Added Command Line Interface (model_swap_agent.py):

  • Full CLI support with model selection
  • Support for all model options
  • Resolution and search grounding options for Nano Banana Pro
  • Verbose output and error handling

• Multi-Model Support:

  • Nano Banana: Fast 1024px generation
  • Nano Banana Pro: High-quality 4K resolution (default)
  • FLUX 2 Pro: Professional quality with custom dimensions
  • FLUX 2 Flex: Advanced controls (guidance, steps)

• Documentation:

  • Comprehensive Docusaurus documentation (docs/docs/agents/model-swap-agent.md)
  • Updated README.md with Model Swap Agent section
  • Updated sidebars.ts to include new documentation

• Features:

  • Intelligent attribute extraction from natural language prompts
  • Perfect outfit preservation (clothing, colors, patterns, textures)
  • Multi-LLM support (OpenAI, Anthropic, Google)
  • High-resolution output (up to 4K)
  • Professional photography quality maintenance

Testing

• I have tested these changes locally
• I have added/updated tests that prove my fix is effective or my feature works
• New and existing unit tests pass locally with my changes

Test Environment

• Python Version: 3.10+
• OS: macOS / Linux
• GPU: N/A (uses cloud APIs)

Test Cases

1. ✅ Basic model swap with Nano Banana Pro (default)
2. ✅ Model swap with FLUX 2 Pro
3. ✅ Model swap with FLUX 2 Flex
4. ✅ Model swap with Nano Banana (fast)
5. ✅ Resolution options for Nano Banana Pro (1K, 2K, 4K)
6. ✅ Search grounding for Nano Banana Pro
7. ✅ Different LLM providers (OpenAI, Anthropic, Google)
8. ✅ Command line interface with all options
9. ✅ Error handling and validation
10. ✅ Outfit preservation verification

Checklist

• My code follows the project's style guidelines
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published

Screenshots / Examples

Command Line Usage

# Basic usage with default Nano Banana Pro
python model_swap_agent.py \
    --image model.jpg \
    --prompt "Replace with a professional male model in his 30s, athletic build"

# Use FLUX 2 Pro for high-quality results
python model_swap_agent.py \
    --image model.jpg \
    --prompt "Replace with a professional female model" \
    --model flux2_pro

# 4K professional quality
python model_swap_agent.py \
    --image outfit.jpg \
    --prompt "Asian female model, mid-20s, athletic, confident pose" \
    --model nano_banana_pro \
    --resolution 4K

Python API Usage

from tryon.agents.model_swap import ModelSwapAgent

# Initialize agent
agent = ModelSwapAgent(llm_provider="openai")

# Generate model swap
result = agent.generate(
    image="person_wearing_outfit.jpg",
    prompt="Replace with a professional Asian female model in her 30s, athletic build"
)

if result["status"] == "success":
    for idx, image in enumerate(result['images']):
        image.save(f"swapped_model_{idx}.png")

Additional Notes

• The agent uses LangChain 1.x API (create_agent)
• Supports multiple LLM providers for intelligent prompt analysis
• All models preserve outfit details automatically
• Default model is Nano Banana Pro for best quality
• Documentation is comprehensive and includes examples for all use cases
• The implementation follows the same patterns as the existing Virtual Try-On Agent

API Keys Required

• LLM Provider: OPENAI_API_KEY, ANTHROPIC_API_KEY, or GOOGLE_API_KEY
• Nano Banana Models: GEMINI_API_KEY
• FLUX 2 Models: BFL_API_KEY

Model Comparison

Model	Resolution	Speed	Quality	Best For
Nano Banana	1024px	Fast ⚡	Good	Quick iterations
Nano Banana Pro	Up to 4K	Medium	Excellent	Professional e-commerce (default)
FLUX 2 Pro	Custom	Medium	High	Professional quality
FLUX 2 Flex	Custom	Slower	Highest	Fine-tuned control

[github-actions]

PR Compliance Checks Passed!