Add Requesty.ai as a new provider option, enabling access to 300+ AI models
through a unified API. This point release focuses on expanding model access
and reducing costs while maintaining full compatibility.

New Features:
- Requesty.ai provider support with 300+ models
- OpenAI models (GPT-4o, GPT-4o-mini, GPT-4-turbo) via Requesty
- DeepSeek Chat with native tool calling support
- Automatic max_tokens optimization (capped at 8192)
- Claude Code wrapper integration for seamless SDK usage
- Full --verbose flag support for debugging

Documentation:
- Updated README.md with Requesty configuration
- Added .env setup instructions
- Documented supported models and pricing
- Added usage examples for agents and Claude Code mode

Benefits:
- 95% cost savings vs Anthropic API
- 300+ model options from single API key
- Native tool calling for GPT-4o and DeepSeek
- Full MCP tool integration (213 tools)

Configuration:
- Environment variable: REQUESTY_API_KEY
- CLI flag: --provider requesty
- Auto-enable: USE_REQUESTY=true
- Default model: deepseek/deepseek-chat

Commits included:
- 731d9b3 feat: Add Requesty.ai provider integration
- f04988d feat: Add Requesty support to Claude Code wrapper

🤖 Generated with Claude Code

Co-Authored-By: Claude <noreply@anthropic.com>