Quick Start Guide
Get up and running with Radium in minutes. This guide will walk you through creating your first agent and running your first orchestrated workflow.
Prerequisitesβ
- Radium installed (Installation Guide)
- At least one AI provider API key:
- Google AI API key (for Gemini)
- Anthropic API key (for Claude)
- OpenAI API key (for GPT models)
Step 1: Set Up Your API Keyβ
# Set your API key (choose one)
export GOOGLE_AI_API_KEY="your-key-here"
# or
export ANTHROPIC_API_KEY="your-key-here"
# or
export OPENAI_API_KEY="your-key-here"
Add this to your shell profile (~/.bashrc, ~/.zshrc, etc.) to persist it.
Step 2: Initialize a Workspaceβ
# Navigate to your project directory
cd my-project
# Initialize Radium workspace
rad init
This creates a .radium/ directory with workspace configuration.
Step 3: Create Your First Agentβ
Create a simple agent configuration:
# Create agent directory structure
mkdir -p agents/my-first-agent
# Create agent configuration
cat > agents/my-first-agent/agent.toml << 'EOF'
[agent]
id = "my-first-agent"
name = "My First Agent"
description = "A simple agent for getting started"
prompt_path = "prompts/my-first-agent.md"
engine = "gemini"
model = "gemini-2.0-flash-exp"
EOF
# Create a simple prompt
mkdir -p prompts
cat > prompts/my-first-agent.md << 'EOF'
# My First Agent
You are a helpful coding assistant. Help users write, debug, and understand code.
## Instructions
- Provide clear, concise explanations
- Write well-documented code
- Suggest best practices
EOF
Step 4: Test Your Agentβ
# List your agents
rad agents list
# Run a simple command
rad step my-first-agent "Write a hello world function in Python"
Step 5: Use Orchestration (TUI)β
For a more interactive experience, use the TUI with orchestration:
# Start the TUI
rad tui
# In the TUI, just type naturally:
# "I need help writing a REST API endpoint"
The orchestrator will automatically:
- Analyze your request
- Select the appropriate agent
- Execute the task
- Return results
Step 6: Create a Planβ
Generate a structured plan from a natural language specification:
# Create a simple spec file
cat > my-feature.md << 'EOF'
# Build a REST API
Create a REST API with:
- User authentication
- CRUD operations for todos
- JSON responses
EOF
# Generate a plan
rad plan my-feature.md
# Execute the plan
rad craft REQ-001
Next Stepsβ
- Core Concepts - Understand agents, orchestration, and workflows
- Configuration Guide - Customize Radium settings
- User Guide - Learn advanced features
- Agent Configuration - Create sophisticated agents
Common First Tasksβ
Create a Code Review Agentβ
rad agents create code-reviewer "Code Reviewer" \
--prompt "You are a senior code reviewer. Review code for bugs, security issues, and best practices."
Set Up Orchestrationβ
Orchestration is enabled by default in TUI. Configure it:
# View orchestration config
cat ~/.radium/orchestration.toml
# Or use TUI commands:
# /orchestrator - Show status
# /orchestrator switch gemini - Change provider
Create an Extensionβ
Package your agent as an extension:
rad extension create my-extension \
--author "Your Name" \
--description "My custom extension"
Troubleshootingβ
Agent Not Foundβ
# Verify agent is discovered
rad agents list
# Check agent configuration
rad agents info my-first-agent
# Validate configuration
rad agents validate
API Key Issuesβ
# Check authentication status
rad auth status
# Test with a specific provider
rad engines health gemini
TUI Not Startingβ
# Check if server is running
rad status
# Start server manually if needed
rad server start
Learn Moreβ
- Installation - Detailed installation instructions
- Core Concepts - Deep dive into Radium concepts
- User Guide - Complete user documentation
- Roadmap - See where Radium is heading
Ready for more? Check out the User Guide to explore advanced features like the persona system, vibe check, and learning system.