Quickstart

Get LunaOS running in under 2 minutes.

Prerequisites

• Node.js 18+ (we recommend using Volta or nvm)
• A LunaOS account — Sign up free

Step 1: Install the CLI

npm install -g @luna-agents/cli

Verify the installation:

luna --version
# @luna-agents/cli v0.2.0

Step 2: Authenticate

luna login

This opens your browser for authentication. After login, the CLI stores your token locally.

Alternatively, use an API key:

# Generate a key from the dashboard or API
curl -X POST https://api.lunaos.ai/api-keys \
  -H "Authorization: Bearer YOUR_JWT_TOKEN" \
  -d '{"name": "my-cli-key"}'

# Set it in the CLI
export LUNAOS_API_KEY="lnos_live_..."

Step 3: Run Your First Agent

luna run code-review --context "
function getUserById(id) {
  const query = 'SELECT * FROM users WHERE id = ' + id;
  return db.execute(query);
}
"

Expected output:

🔍 Code Review Agent — Analyzing...

## Security Issues

### 🔴 Critical: SQL Injection Vulnerability
**Line 2**: String concatenation in SQL query allows injection attacks.

**Fix:**
\`\`\`javascript
function getUserById(id) {
  const query = 'SELECT * FROM users WHERE id = ?';
  return db.execute(query, [id]);
}
\`\`\`

### 🟡 Warning: No Input Validation
The `id` parameter is not validated before use.

**Fix:** Add type checking and sanitization.

---
📊 Execution: 1.2s | Tokens: 847 | Cost: $0.0003

Step 4: Try More Agents

# Generate tests for a file
luna run testing-validation --context "$(cat src/auth.ts)"

# Write documentation
luna run documentation --context "$(cat src/api/routes.ts)"

# Analyze requirements
luna run requirements-analyzer --context "Build a real-time chat application"

Step 5: Use Agent Chains

Chain multiple agents together for complex workflows:

luna chain run code-review-pipeline --context "$(cat src/main.ts)"

Or define custom chains:

{
  "name": "Full Review",
  "steps": [
    { "agent": "code-review", "passOutput": true },
    { "agent": "testing-validation", "passOutput": true },
    { "agent": "documentation" }
  ]
}

Using the REST API Directly

If you prefer to integrate programmatically:

# Get a token
TOKEN=$(curl -s -X POST https://api.lunaos.ai/auth/login \
  -H "Content-Type: application/json" \
  -d '{"email":"you@example.com","password":"your-password"}' \
  | jq -r '.token')

# Run an agent (SSE stream)
curl -N -X POST https://api.lunaos.ai/agents/execute \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "agent": "code-review",
    "context": "your code here",
    "provider": "deepseek",
    "model": "deepseek-chat"
  }'

What's Next?

• Agent Catalog → — Browse all 28 agents with examples
• API Reference → — Full REST API documentation
• Configuration → — Customize your setup