Skip to main content

Direct API Access

Overview

Codebolt exposes all the major functions needed to create an AI Agent using the CodeboltJS SDK. The SDK provides a comprehensive set of functions organized into logical categories for creating powerful AI agents. Although these functions can be used in any mode, they are primarily used for Agent creation in Direct Function Mode.

API Categories

The Codebolt API is organized into the following categories:

Core AI & Machine Learning

Advanced AI capabilities for multi-agent systems, knowledge management, and memory:

  • Swarm - Multi-agent coordination with teams, roles, and vacancies
  • Knowledge Graph - Graph-based knowledge with entities and relationships
  • Episodic Memory - Time-series contextual experiences
  • Persistent Memory - Long-term knowledge storage
  • Memory Ingestion - Automated knowledge base building

Development & Tools

Tools for development, search, and integration:

  • Search - Advanced codebase search
  • Crawler - Web crawling and data extraction
  • MCP - Model Context Protocol for modular tools
  • State - Application state management
  • Utils - General utility functions

Agent & Task Management

Create and manage AI agents and their tasks:

  • Agent - Customized agent creation
  • Agent Portfolio - Reputation and skill tracking
  • Agent Deliberation - Decision-making capabilities
  • Action Block - Structured workflows
  • Auto Testing - Automated test management
  • Capability - Agent skill definitions

Communication & Collaboration

Team communication and feedback systems:

  • Thread - Conversation management
  • Mail - Email integration
  • Hook - Webhook management
  • Group Feedback - Collaborative feedback
  • Chat - Real-time messaging

Project & Planning

Project management and planning tools:

  • Calendar - Event scheduling and management
  • Requirement Plan - Requirements tracking
  • Review Merge Request - Code review workflow
  • Roadmap - Project milestones
  • Task Planner - Task orchestration

Data & Storage

Persistent data management:

  • KV Store - Key-value storage with querying
  • Event Log - Activity logging
  • File Update Intent - Change tracking
  • Context Assembly - Multi-source context

Context & Analysis

Dynamic context and task management:

  • Context Rule Engine - Dynamic context selection
  • Todo - Task management
  • Side Execution - Isolated process execution

User & Message Management

User communication handling:

  • User Message Manager - Message workflows
  • User Message Utilities - Message processing

Core Infrastructure

Essential system modules:

  • Browser - Web automation
  • Fs - File operations
  • Git - Version control
  • Terminal - Command execution
  • LLM - Language model integration
  • Vector DB - Embedding storage

All available modules:

  • Swarm - Multi-agent swarm coordination with teams, roles, and vacancies for orchestrating complex AI systems
  • Knowledge Graph - Graph-based knowledge management with entities, relationships, and pattern matching queries
  • Episodic Memory - Time-series episodic memory for storing and retrieving contextual experiences
  • Persistent Memory - Long-term persistent memory storage for agent knowledge and experiences
  • Memory Ingestion - Automated memory ingestion and processing for building agent knowledge bases
  • Search - Advanced codebase search with semantic and pattern matching capabilities
  • Crawler - Web crawling and content extraction for gathering external data
  • MCP - Model Context Protocol for managing modular tools and services through WebSocket communication
  • State - Application state management with real-time updates and persistence
  • Utils - General utility functions and helpers for common operations
  • Agent - Create customized, high-performance agents tailored to specific needs
  • Agent Portfolio - Agent reputation system with karma, testimonials, talents, and rankings
  • Agent Deliberation - Agent decision-making and reasoning capabilities for complex tasks
  • Action Block - Structured action blocks for organizing and executing agent workflows
  • Auto Testing - Automated testing with test suites, cases, and execution tracking
  • Capability - Define and manage agent capabilities and skill sets
  • Thread - Conversation thread management with messages and file change tracking
  • Mail - Email management and communication capabilities
  • Hook - Webhook management for event-driven integrations
  • Group Feedback - Collaborative feedback collection and management for teams
  • Chat - Real-time chat interface with message history and user interaction
  • Calendar - Event management with scheduling, reminders, and recurring events
  • Requirement Plan - Project requirement planning and tracking
  • Review Merge Request - Merge request review and code collaboration
  • Roadmap - Project roadmap planning and milestone tracking
  • Task Planner - Task planning and execution orchestration
  • KV Store - Persistent key-value storage with namespaces and querying capabilities
  • Event Log - Event logging and audit trail for system activities
  • File Update Intent - Track and manage file modification intentions and changes
  • Context Assembly - Assemble and manage context from multiple memory sources
  • Context Rule Engine - Dynamic context selection with conditional rules and prioritization
  • Todo - Task management with priorities, tags, and import/export capabilities
  • Side Execution - Execute code in isolated processes for long-running or intensive tasks
  • User Message Manager - Manage user messages and communication workflows
  • User Message Utilities - Utility functions for processing and handling user messages
  • Browser - Browser automation for web interaction, scraping, and testing
  • Fs - File system operations for reading, writing, and managing files
  • Git - Git version control operations for repository management
  • Tool - Tool management and execution framework
  • Cbstate - Application state tracking with continuous updates
  • Code Utils - Code parsing and manipulation utilities
  • Code Parsers - AST generation and code structure analysis utilities
  • Job - Job management and execution tracking
  • Output Parsers - Parse and structure output from various sources
  • Project - Project management and configuration
  • Terminal - Terminal command execution with support for long-running processes
  • Codebase Search - Search and navigate codebases with advanced queries
  • Codemap - Visual code mapping and structure analysis
  • Project Structure - Analyze and manage project structure and architecture
  • Project Structure Update Request - Multi-agent coordination for proposing and reviewing structural changes
  • History - Chat history management with summarization
  • Db Memory - Database-backed memory for persistent knowledge storage
  • Debug - Debugging utilities and logging capabilities
  • LLM - Large Language Model integration and management
  • RAG - Retrieval-Augmented Generation for enhanced AI responses
  • Tokenizer - Text tokenization for processing and analysis
  • Vector DB - Vector database for embedding storage and similarity search

Quick Start

import codebolt from '@codebolt/codeboltjs';

// Initialize connection
await codebolt.waitForReady();

// Use any module
const result = await codebolt.swarm.createSwarm({
name: 'My Team',
description: 'A collaborative agent swarm'
});