- Synchronized ImageGen node with 300s timeout and forced REST mode
- Added model mapping for ImageGen to use gemini-2.0-flash by default
- Updated README with specific steps to fix 'Limit: 0' quota errors
- Switch from generate_content to generate_content_stream for text generation
- Disable Automatic Function Calling (AFC) to reduce overhead
- Increase client timeout to 300s (5 min) for complex prompts
- Add chunk/character count logging for streaming responses