API Reference#
🚧 Early Access Documentation
This documentation is part of an early access release and is under active development. Many sections are still being written, edited, or reorganized. Expect inconsistencies, missing content, outdated references, and broken cross-links.
We welcome feedback! If you find issues or have suggestions, please open an issue on our GitHub page.
Complete technical reference for all public APIs in the Alpha Berkeley Framework. This is your authoritative source for classes, methods, functions, and their precise interfaces. See also the Developer Guides for learning-oriented content, then return here when you need specific implementation details.
🎯 API Learning Paths & Quick Navigation
Essential API Learning Path
Base Components - BaseCapability and decorators
State and Context Management - AgentState and ContextManager
Gateway - Entry point architecture
Human Approval - Safety and oversight
Classification System - Robust error handling
Core API Patterns:
framework.base.BaseCapability
with@capability_node
decoratorframework.state.AgentState
for LangGraph-native state managementframework.context.ContextManager
for type-safe data exchangeframework.registry.RegistryManager
for component discovery
Advanced API Workflows
Orchestration - ExecutionPlan creation and coordination
Python Execution - Secure code execution APIs
Data Management - Unified data source integration
Container Management - Deployment and service management
Recovery Coordination - Recovery strategies and coordination
Production-Ready APIs:
framework.approval.ApprovalManager
- LangGraph-native approval workflowsframework.data_management.DataSourceManager
- Provider discovery and retrievaldeployment.container_manager
- Service orchestration and deploymentframework.models.get_chat_completion()
- Multi-provider LLM managementConfiguration System - Environment and settings management
Framework Utilities - Logging, streaming, and observability
Implementation Shortcuts
Build a new capability → BaseCapability
+ capability_node()
Add approval workflows → ApprovalManager
+ Human Approval
Execute Python code safely → PythonExecutorService
Store user context → ContextManager
+ State and Context Management
Deploy with containers → container_manager
+ Container Management
Handle complex data flows → DataSourceManager
Integrate external data → DataSourceProvider
+ provider patterns
Manage LLM models → get_model()
+ get_chat_completion()
Configure logging → get_logger()
+ ComponentLogger
Stream responses → get_streamer()
+ StreamWriter
Core Framework APIs
Base Components - BaseCapability, BaseInfrastructureNode, decorators
State and Context Management - AgentState, StateManager, ContextManager
Registry System - RegistryManager, component discovery
Configuration System - Environment and settings management
Prompt System - Framework prompt providers and builders
Infrastructure Pipeline APIs
Gateway - Gateway class, message processing, state lifecycle
Task Extraction - TaskExtractionNode, conversation analysis
Classification - ClassificationNode, capability selection
Orchestration - OrchestrationNode, ExecutionPlan creation
Execution Control - Router coordination and flow management
Message Generation - Response generation and clarification
Production & Error Management
Human Approval - ApprovalManager, approval workflows
Data Management - DataSourceManager, provider framework
Python Execution - PythonExecutorService, secure execution
Memory Storage - Memory management and persistence
Container Management - Container orchestration and deployment
Classification System - Error classification and severity levels
Exception Reference - Exception hierarchy and handling
Recovery Coordination - Recovery strategies and coordination
Framework Utilities - Logging, streaming, and developer tools
Core API Sections#
Essential daily-use APIs
State management, context handling, base components, registry system, and configuration APIs that developers use most frequently.
Message processing pipeline
Gateway, task extraction, classification, orchestration, and execution control for building intelligent workflows.
Production-ready services
Container management, data management, human approval systems, and deployment tools for building robust applications.
Resilient error management
Exception hierarchy, recovery strategies, and error classification for building fault-tolerant systems.
Advanced utilities
Model factory, logging, streaming, and developer tools for extending and customizing framework behavior.
📚 Complete API Documentation Tree
🏗️ Core Framework
Base Components - Foundation classes, decorators, and planning
State and Context Management - LangGraph state management and tracking
Registry System - Component discovery and registration
Configuration System - Configuration management
Prompt System - Framework prompt providers and builders
⚡ Infrastructure
Gateway - Main entry point and request routing
Task Extraction - User input analysis and task identification
Classification - Capability selection and routing logic
Orchestration - Execution planning and coordination
Message Generation - Response formatting and user communication
Execution Control - Execution flow and state management
🚀 Production Systems
Human Approval - LangGraph-native approval workflows
Data Management - Unified data source integration
Python Execution - Secure code generation and execution
Memory Storage - Persistent user memory management
Container Management - Service orchestration and deployment
🛡️ Error Handling
Classification System - Structured exception classification
Exception Reference - Exception hierarchy and reference
Recovery Coordination - Intelligent error recovery and retry
🔧 Framework Utilities
Framework Utilities - Model factory, logging and streaming utilities