Core Overview

MyDeskBot is a powerful AI-assisted development platform that provides comprehensive development support.

Core Concepts

AI Assistant

The core of MyDeskBot is an intelligent AI assistant that can:

• 🧠 Understand Code - Deeply understand your codebase
• 📝 Generate Code - Intelligently generate high-quality code
• 🔍 Code Review - Real-time code quality checking
• 💡 Provide Suggestions - Recommendations based on best practices
• 🎯 Context Awareness - Understand the entire project context

Model Providers

MyDeskBot supports multiple AI model providers:

• OpenAI - GPT-4, GPT-3.5 Turbo
• Anthropic - Claude 3 Opus, Sonnet, Haiku
• Google - Gemini Pro
• Open Source Models - Use local models through Ollama

Tools and Plugins

Rich ecosystem of tools and plugins:

• MCP Tools - Model Context Protocol integration
• IDE Plugins - Integration with your favorite editors
• Desktop Application - Standalone AI assistant application
• Command Line Tools - AI assistance in the terminal

Architecture Overview

┌─────────────────────────────────────────────────┐
│               User Interface                     │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐     │
│  │ IDE Plugin│  │ Desktop  │  │ CLI Tool │     │
│  │           │  │  App     │  │          │     │
│  └──────────┘  └──────────┘  └──────────┘     │
└─────────────────────────────────────────────────┘
                       │
                       ▼
┌─────────────────────────────────────────────────┐
│               MyDeskBot Core                     │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐     │
│  │ Context  │  │ Prompt   │  │ Rules    │     │
│  │ Engine   │  │ System   │  │ System   │     │
│  └──────────┘  └──────────┘  └──────────┘     │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐     │
│  │ Model    │  │ Cache    │  │ Error    │     │
│  │ Adapter  │  │ Manager  │  │ Handler  │     │
│  └──────────┘  └──────────┘  └──────────┘     │
└─────────────────────────────────────────────────┘
                       │
                       ▼
┌─────────────────────────────────────────────────┐
│              Model Providers                     │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐     │
│  │  OpenAI  │  │ Anthropic│  │ Other    │     │
│  │          │  │          │  │ Providers│     │
│  └──────────┘  └──────────┘  └──────────┘     │
└─────────────────────────────────────────────────┘

Main Components

1. Context Engine

Responsible for understanding and analyzing your codebase:

graph LR
    A[Code Scanning] --> B[Syntax Analysis]
    B --> C[Symbol Extraction]
    C --> D[Dependency Analysis]
    D --> E[Pattern Recognition]
    E --> F[Knowledge Graph]

Features:

• Index codebase
• Understand code structure
• Identify code patterns
• Maintain dependency relationships

2. Prompt System

Intelligently generates and manages prompts:

Features:

• Automatic prompt generation
• Prompt template management
• Context injection
• Prompt optimization

3. Rules System

Configurable rules and constraints:

Features:

• Code style rules
• Security rules
• Performance rules
• Custom rules

4. Model Adapter

Unified model interface:

Supports:

• OpenAI API
• Anthropic API
• Azure OpenAI
• Local models (Ollama)
• Custom endpoints

5. MCP Tool Integration

Model Context Protocol tool integration:

Integrated Tools:

• GitHub
• Notion
• Supabase
• Sentry
• Netlify
• And more...

Workflows

1. Code Completion Workflow

User Input
    ↓
Context Collection
    ↓
Prompt Generation
    ↓
Model Inference
    ↓
Result Filtering
    ↓
Display Suggestions

2. Code Review Workflow

Select Code
    ↓
Static Analysis
    ↓
Rule Checking
    ↓
AI Review
    ↓
Generate Report
    ↓
Display Suggestions

3. Plan Mode Workflow

Describe Task
    ↓
Requirement Analysis
    ↓
Generate Plan
    ↓
User Confirmation
    ↓
Step-by-step Execution
    ↓
Result Verification

Core Features

Intelligent Awareness

• Project Awareness - Understands entire project structure
• Pattern Learning - Learns your coding style
• Context Relevance - Based on the entire codebase

Highly Configurable

• Model Selection - Supports multiple AI models
• Custom Rules - Configurable rule system
• Plugin Extension - Rich plugin ecosystem

Privacy & Security

• Local Processing - Sensitive operations can be done locally
• Data Encryption - Encrypted transmission and storage
• Controlled Scope - Control what data is sent

High Performance

• Intelligent Caching - Reduces duplicate requests
• Streaming Output - Fast response to users
• Incremental Indexing - Efficient codebase indexing

Use Cases

Daily Development

• Code completion
• Quick fixes
• Code explanation
• Documentation generation

Code Review

• Quality checking
• Security review
• Performance optimization
• Best practices

Learning Assistance

• Technical learning
• Code examples
• Concept explanation
• Best practices

Automation

• Script generation
• CI/CD configuration
• Test generation
• Documentation automation

Technology Stack

Backend

• Node.js - Runtime
• TypeScript - Primary language
• Rust - Performance-critical components

Frontend

• React - UI framework
• Electron - Desktop application
• Tauri - Next-generation desktop application

AI/ML

• OpenAI API - Model services
• Anthropic API - Claude models
• LangChain - AI application framework

Extensibility

Plugin System

• IDE plugins (IntelliJ, VS Code, Neovim)
• MCP tool integration
• Custom rules

API Integration

• REST API
• GraphQL API
• WebSocket (real-time communication)

Local Deployment

• Docker support
• Kubernetes support
• Private deployment

Configuration Management

Configuration File

# config.yaml
name: My MyDeskBot Configuration
version: 0.0.1
schema: v1

models:
  - name: "gpt-4"
    provider: "openai"
    model: "gpt-4"
    apiKey: "${{ secrets.OPENAI_API_KEY }}"
    roles:
      - chat
      - edit
      - apply

  - name: "claude-3-sonnet"
    provider: "anthropic"
    model: "claude-3-sonnet"
    apiKey: "${{ secrets.ANTHROPIC_API_KEY }}"
    roles:
      - chat
      - autocomplete

  - name: "local-llama"
    provider: "ollama"
    model: "llama2"
    apiBase: "http://localhost:11434"
    roles:
      - chat

Environment Variables

# ~/.mydeskbot/.env
OPENAI_API_KEY=your-openai-key
ANTHROPIC_API_KEY=your-anthropic-key