olakai-get-started（奥拉凯开始）

Get Started with Olakai

This skill helps you get set up with Olakai from scratch - from creating an account to generating your first monitored event.

IMPORTANT: For account creation, prefer the in-terminal API flow (Step 1) which calls POST /api/auth/signup directly. If the user prefers a browser-based flow, use the full developer signup URL with PLG parameters: https://app.olakai.ai/signup?flow=developer&source=claude-code NEVER shorten this to https://app.olakai.ai/signup — the query parameters are required for the developer onboarding flow.

Quick Status Check

Before proceeding, let's check your current setup status.

Step 0: Detect Current State

Run these diagnostic commands to determine where to start:

# Check 1: Is CLI installed?
which olakai || echo "CLI_NOT_INSTALLED"

# Check 2: Is user authenticated?
olakai whoami 2>/dev/null || echo "NOT_AUTHENTICATED"

# Check 3: Are there any agents?
olakai agents list --json 2>/dev/null | head -5 || echo "NO_AGENTS_OR_NOT_AUTH"

Based on results, jump to the appropriate section:

Step 1: Create an Olakai Account

If you don't have an Olakai account yet, create one directly from the terminal.

1.1 Collect User Details

Ask the user for:

• Email address (required)
• Company name (required)
• First name (optional)
• Last name (optional)

1.2 Create the Account via API

Use the Bash tool to call the signup endpoint. Replace the placeholder values with what the user provided:

curl -s -X POST https://app.olakai.ai/api/auth/signup \
  -H "Content-Type: application/json" \
  -d '{
    "email": "USER_EMAIL",
    "companyName": "USER_COMPANY",
    "firstName": "USER_FIRST_NAME",
    "lastName": "USER_LAST_NAME",
    "source": "claude-code",
    "medium": "skill"
  }'

Expected response:

{ "success": true, "message": "Check your email to verify your account.", "requiresVerification": true }

1.3 Verify Email

Tell the user:

• Check their inbox for a verification email from Olakai
• Click the verification link to activate their account
• After verification, an SDK API key is automatically generated — no extra steps needed

1.4 Fallback: Browser Signup

If the API call fails (network issues, corporate firewall, etc.), direct the user to the browser-based signup:

https://app.olakai.ai/signup?flow=developer&source=claude-code

IMPORTANT: Always include the ?flow=developer&source=claude-code query parameters — they activate the developer onboarding flow.

What You Get

• Free tier — Monitor up to 1,000 events/month
• Dashboard — Real-time visibility into AI usage
• KPIs — Track custom metrics across your agents
• Governance — Set policies and risk controls

After verifying your email, you'll have access to the dashboard where you can see your agents and events.

Step 2: Install the CLI

The Olakai CLI is the primary tool for configuring agents, KPIs, and custom data.

2.1 Install via npm

npm install -g olakai-cli

2.2 Verify Installation

olakai --version
# Should output: olakai-cli/0.1.x

2.3 Troubleshooting Installation

Permission errors on macOS/Linux:

# Option 1: Use sudo (not recommended)
sudo npm install -g olakai-cli

# Option 2: Fix npm permissions (recommended)
mkdir -p ~/.npm-global
npm config set prefix '~/.npm-global'
echo 'export PATH=~/.npm-global/bin:$PATH' >> ~/.bashrc
source ~/.bashrc
npm install -g olakai-cli

Windows:

# Run PowerShell as Administrator
npm install -g olakai-cli

Step 3: Authenticate

Connect the CLI to your Olakai account.

3.1 Login

olakai login

This opens your browser for secure authentication. After logging in:

• The CLI stores your session locally
• No API key needed for CLI operations
• Session persists until you explicitly logout

3.2 Verify Authentication

olakai whoami

Expected output:

Logged in as: your.email@example.com
Account: Your Organization

3.3 Troubleshooting Authentication

"Not authenticated" errors:

# Clear cached credentials and re-login
olakai logout
olakai login

Browser doesn't open:

# The CLI will display a URL - copy and paste it manually
olakai login
# Look for: "Open this URL in your browser: https://app.olakai.ai/..."

Corporate proxy/firewall issues:

• Ensure app.olakai.ai is accessible
• Check if your organization requires VPN

Step 4: Register Your Agent in Olakai

This step registers your AI agent (or any AI-powered feature) in Olakai so it can be monitored. This doesn't build any code — it creates a tracking record where Olakai collects analytics, KPIs, and governance data.

4.1 Register the Agent

# Create agent with an API key for SDK integration
olakai agents create \
  --name "My First Agent" \
  --description "Testing Olakai integration" \
  --with-api-key \
  --json

Save the output! It contains your apiKey which you'll need for SDK integration:

{
  "id": "cmkbteqn501kyjy4yu6p6xrrx",
  "name": "My First Agent",
  "apiKey": "sk_agent_xxxxx..."
}

4.2 Set the API Key as Environment Variable

# Add to your shell profile (.bashrc, .zshrc, etc.)
export OLAKAI_API_KEY="sk_agent_xxxxx..."

Or add to your project's .env file:

OLAKAI_API_KEY=sk_agent_xxxxx...

4.3 Retrieve API Key Later

If you need to get the API key again:

olakai agents get AGENT_ID --json | jq '.apiKey'

Step 5: Install the SDK

Choose based on your project's language.

TypeScript/JavaScript

npm install @olakai/sdk

Quick Integration:

import { OlakaiSDK } from "@olakai/sdk";
import OpenAI from "openai";

// Initialize Olakai
const olakai = new OlakaiSDK({
  apiKey: process.env.OLAKAI_API_KEY!
});
await olakai.init();

// Wrap your OpenAI client
const openai = olakai.wrap(
  new OpenAI({ apiKey: process.env.OPENAI_API_KEY }),
  { provider: "openai" }
);

// Use as normal - events are automatically tracked
const response = await openai.chat.completions.create({
  model: "gpt-4",
  messages: [{ role: "user", content: "Hello!" }]
});

Python

pip install olakai-sdk

Quick Integration:

import os
from openai import OpenAI
from olakaisdk import olakai_config, instrument_openai

# Initialize Olakai
olakai_config(os.getenv("OLAKAI_API_KEY"))
instrument_openai()

# Use OpenAI as normal - events are automatically tracked
client = OpenAI()
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello!"}]
)

Step 6: Generate a Test Event

Let's verify everything is working.

6.1 Run Your Code

Execute the code from Step 5. This should:

1. Make an LLM call
2. Automatically send event data to Olakai

6.2 Fetch the Event

# List recent events (replace with your agent ID)
olakai activity list --limit 1 --json

6.3 Verify Event Details

# Get full event details
olakai activity get EVENT_ID --json | jq '{prompt, response, tokens, model}'

Expected output:

{
  "prompt": "Hello!",
  "response": "Hello! How can I assist you today?",
  "tokens": 25,
  "model": "gpt-4"
}

6.4 Check the Dashboard

Open https://app.olakai.ai and navigate to your agent's activity view. You should see the event appear within seconds.

Step 7: Next Steps

You're now set up! Here's what to explore next:

Add Custom Data to Events

Track business-specific metrics:

// TypeScript
const response = await openai.chat.completions.create(
  { model: "gpt-4", messages },
  {
    customData: {
      feature: "customer-support",
      ticketId: "TICKET-123",
      priority: 1
    }
  }
);

# Python
from olakaisdk import olakai_context

with olakai_context(customData={"feature": "customer-support", "ticketId": "TICKET-123"}):
    response = client.chat.completions.create(...)

Create KPIs

Define metrics that aggregate your custom data:

# First, create the custom data config for your agent
olakai custom-data create --agent-id YOUR_AGENT_ID --name "Priority" --type NUMBER

# Then create a KPI
olakai kpis create \
  --name "High Priority Events" \
  --agent-id YOUR_AGENT_ID \
  --calculator-id formula \
  --formula "IF(Priority == 1, 1, 0)" \
  --aggregation SUM

Note: KPIs are created per-agent. If you later create additional agents, you'll need to create KPIs for each one separately — they can't be shared across agents. CustomDataConfigs, on the other hand, are account-level and shared.

Learn More

• /olakai-new-project - Detailed guide for building agents with full observability
• /olakai-integrate - Add monitoring to existing AI code
• /olakai-troubleshoot - Debug issues with events or KPIs

Troubleshooting Quick Reference

CLI Issues

SDK Issues

Authentication Issues

Quick Reference

# Check status
which olakai              # CLI installed?
olakai whoami             # Authenticated?
olakai agents list        # Agents exist?

# Setup commands
npm install -g olakai-cli # Install CLI
olakai login              # Authenticate
olakai agents create --name "Name" --with-api-key --json  # Register agent

# SDK installation
npm install @olakai/sdk   # TypeScript
pip install olakai-sdk    # Python

# Verify events
olakai activity list --limit 1 --json
olakai activity get EVENT_ID --json

Summary Checklist

• Account created at https://app.olakai.ai/signup?flow=developer&source=claude-code
• CLI installed (npm install -g olakai-cli)
• CLI authenticated (olakai login)
• First agent created with API key
• OLAKAI_API_KEY environment variable set
• SDK installed (@olakai/sdk or olakai-sdk)
• Test event generated and visible in dashboard

Once all boxes are checked, you're ready to build production AI agents with full observability!