# What is RiceCoder?

**Last Updated**: December 3, 2025

RiceCoder (`rice`) is a **terminal-first, spec-driven coding assistant** that helps developers write better code through research, analysis, and intelligent code generation.

## Overview

RiceCoder combines a research-first approach with spec-driven development methodology to help you write better code systematically. Instead of generating code immediately, RiceCoder analyzes your project, understands your patterns, and generates code that fits your style. This guide explains what RiceCoder is, how it works, and why you should use it.

## Core Philosophy

Unlike traditional AI coding tools that generate code immediately, RiceCoder follows a **research-first approach**:

1. **Research** - Analyze your project structure, patterns, and context
2. **Understand** - Build a comprehensive understanding of your codebase
3. **Generate** - Create code that fits your project's style and patterns
4. **Validate** - Ensure generated code meets quality standards

## Key Features

### 🔬 Research-First Analysis
- Automatic project structure detection
- Language and framework identification
- Dependency analysis
- Pattern recognition
- Context-aware code generation

### 📋 Spec-Driven Development
- Define requirements with acceptance criteria
- Design solutions systematically
- Generate implementation tasks
- Track progress with clear milestones
- Maintain traceability from requirements to code

### 💻 Terminal-Native Interface
- Beautiful CLI with colored output
- Interactive TUI (Terminal User Interface)
- Works over SSH and remote connections
- No GUI dependencies
- Keyboard-first navigation

### 🔒 Privacy-First & Offline-Capable
- Local model support via Ollama
- Offline-first architecture
- No data sent to cloud unless you choose
- Full control over your code and context

### 🤖 Multi-Agent Framework
- Specialized agents for different tasks
- Code review agent
- Test generation agent
- Documentation agent
- Extensible agent system

### 🎨 Multi-Provider Support
- OpenAI (GPT-4, GPT-3.5)
- Anthropic (Claude)
- GitHub Copilot
- Local models (Ollama)
- 75+ provider support

## How It Works

### Basic Workflow

```text
1. Initialize Project
   rice init

2. Start Interactive Chat
   rice chat

3. Ask Questions or Request Features
   > Generate a user authentication module

4. RiceCoder:
   - Analyzes your project
   - Understands your patterns
   - Generates code
   - Shows diff for review
   - Applies changes with your approval

5. Iterate
   > Refactor the auth module to use JWT
```

### Spec-Driven Workflow

```text
1. Create Specification
   rice spec create my-feature

2. Define Requirements
   - User stories
   - Acceptance criteria
   - Constraints

3. Design Solution
   - Architecture
   - Data models
   - Integration points

4. Generate Implementation Tasks
   - Hierarchical task breakdown
   - Dependencies
   - Exit criteria

5. Execute Tasks
   - Implement each task
   - Run tests
   - Validate against requirements

6. Complete Feature
   - All tests passing
   - All requirements met
   - Code reviewed
```

## Why Choose RiceCoder?

| Aspect | RiceCoder | GitHub Copilot | Aider | Claude Code |
|--------|-----------|---|---|---|
| **Terminal-Native** | ✅ | ❌ | ✅ | ❌ |
| **Spec-Driven** | ✅ | ❌ | ⚠️ | ❌ |
| **Offline-First** | ✅ | ❌ | ❌ | ❌ |
| **Research-First** | ✅ | ❌ | ⚠️ | ❌ |
| **Multi-Agent** | ✅ | ❌ | ❌ | ⚠️ |
| **Local Models** | ✅ | ❌ | ✅ | ❌ |
| **Open Source** | ✅ | ❌ | ✅ | ❌ |

## Use Cases

### 1. Spec-Driven Development
Perfect for teams that follow specification-driven development methodology. Define requirements, design, and tasks - RiceCoder generates the implementation.

### 2. Code Review & Quality
Use specialized agents to review code, suggest improvements, and ensure quality standards.

### 3. Test Generation
Automatically generate unit tests, integration tests, and property-based tests for your code.

### 4. Documentation
Generate API documentation, README files, and inline code documentation.

### 5. Refactoring
Safely refactor code with impact analysis and automated testing.

### 6. Learning & Onboarding
Analyze unfamiliar codebases and generate explanations and documentation.

## Architecture Overview

RiceCoder is built with a modular, layered architecture:

```text
┌─────────────────────────────────────┐
│         CLI / TUI Interface         │
├─────────────────────────────────────┤
│      Application Layer              │
│  (Commands, Workflows, Sessions)    │
├─────────────────────────────────────┤
│      Domain Layer                   │
│  (Specs, Agents, Providers)         │
├─────────────────────────────────────┤
│      Infrastructure Layer           │
│  (Storage, Files, Config)           │
├─────────────────────────────────────┤
│      External Services              │
│  (AI Providers, Git, LSP)           │
└─────────────────────────────────────┘
```

## Technology Stack

- **Language**: Rust (1.75+)
- **TUI Framework**: Ratatui
- **CLI Framework**: Clap
- **Testing**: Proptest (property-based testing)
- **AI Providers**: OpenAI, Anthropic, Ollama, and more
- **Storage**: YAML/JSON configuration, local file system
- **Version Control**: Git integration

## Getting Started

1. **[Installation & Setup](./Installation-Setup)** - Install RiceCoder
2. **[Quick Start Guide](./Quick-Start)** - Your first project
3. **[Configuration](./Configuration)** - Customize RiceCoder
4. **[CLI Commands](./CLI-Commands)** - Available commands

## Learn More

- **[Spec-Driven Development](./Spec-Driven-Development)** - How to use specs effectively
- **[Architecture Overview](./Architecture-Overview)** - Deep dive into RiceCoder's design
- **[Development Roadmap](./Development-Roadmap)** - What's coming next

## See Also

- [Installation & Setup](./Installation-Setup.md) - Install RiceCoder
- [Quick Start Guide](./Quick-Start.md) - Get started in 5 minutes
- [Configuration Guide](./Configuration.md) - Customize RiceCoder
- [CLI Commands Reference](./CLI-Commands.md) - Available commands
- [Spec-Driven Development Guide](./Spec-Driven-Development.md) - Systematic development methodology

---

*RiceCoder: Think before you code.*

*Last updated: December 3, 2025*