Table of Contents
Table of Contents
- Getting Started
- Core Concepts
- Basic Text Generation
- Direct LLM Access
- Streaming
- Advanced Inference
- Connectors & Authentication
- External API Integration
- Vendor Templates
- OAuth for External APIs
- Scoped Connector Registry
- Enhanced connector execution profiles
- Whole connector pages
- Model Registry
- Endpoint capability metadata
- Effective temperature decisions
- Vendor labels and colors
- Agent Features
- Tools & Function Calling
- Dynamic Tool Management
- Tool outcomes and cancellation
- Tool Permissions
- Tool Execution Plugins
- MCP (Model Context Protocol)
- Context Management
- Selective context assembly
- Experimental context strategies
- Durable tool evidence
- Protected task state
- Managed model routing
- Model tasks and helper models
- Bounded managed-loop policies
- Opt-in Agent presets and inspection
- Unified Store Tools
- InContextMemory (NextGen Plugin)
- Self-Learning Memory (NextGen Plugin)
- Tool Catalog: Dynamic Tool Loading/Unloading
- Shared Workspace
- Session Persistence
- Portable Agent Packages
- Negotiated portable observations
- Async (Non-Blocking) Tools
- Long-Running Sessions (Suspend/Resume)
- Routine Execution
- Verified routines
- Agent Registry
- Agent Orchestrator
- Complete orchestration configuration
- Composed guarantees and inspection
- Agent Runtime (preview)
- Multimodal (Vision)
- Audio (TTS/STT)
- OpenAI Realtime API
- Image Generation
- Embeddings
- Video Generation
- Custom Media Storage
- Web Search
- Web Scraping
- Advanced Features
- Production Deployment
- Examples
- Agentic Evaluation Baselines
- Persistent Instructions (NextGen Plugin)
- User Info (NextGen Plugin)