# Developer Quick Start: Register Your First Agent With Olakai The fastest path to integrating your AI agents with Olakai's analytics, governance, and ROI dashboards. > This guide assumes you are using an AI coding agent such as Claude Code, Cursor, Windsurf, or VS Code Copilot. ## Go Live in 3 mins or less! ### 1) Do you already have an agent to register for Olakai analytics?

Yes Great, navigate to your agent folder and see the next step.

No No problem, if you're just testing, we can help you create an agent with synthetic data for testing: 1 Clone the example repo: ```bash git clone https://github.com/olakai-ai/olakai-agentic-examples.git cd olakai-agentic-examples/examples/ap-invoice-agent ``` 2 Choose TypeScript or Python and install dependencies: **TypeScript:** ```bash cd agent && npm install ``` **Python:** ```bash cd agent-python && pip install -r requirements.txt ``` 3 Run it: ```bash # TypeScript npm start # Python python3 run.py ``` This is a 6-step AP invoice processing agent that processes 10 sample invoices using simulated responses — no API key needed. **Want real LLM traffic? (optional)** Set your OpenAI API key to make live calls to GPT-4o-mini (< $0.01 per run): ```bash cp .env.example .env echo "OPENAI_API_KEY=your-key-here" >> .env ```

### 2) Open your AI coding agent In your project folder, open your AI coding agent: | Agent | How to start | | -------------- | --------------------------------------------------------------------- | | Claude Code | `claude` in terminal | | Cursor | `cursor .` in terminal, or open Cursor and select your folder | | GitHub Copilot | `code .` in terminal (Copilot extension must be installed in VS Code) | | Windsurf | `windsurf .` in terminal, or open Windsurf and select your folder | | Cline | `code .` in terminal (Cline extension must be installed in VS Code) | | Aider | `aider` in terminal | ### 3) Install the Olakai Skill AI coding agents can use Olakai skills to guide you through integration. **Claude Code (Native Plugin)** ```bash # Install the Olakai plugin, inside the interactive Claude code terminal interface /plugin marketplace add olakai-ai/olakai-skills /plugin install olakai@olakai-skills ``` **Other AI Coding Agents (Agent Skills Standard)** | Agent | Installation | | --------------- | ------------------------------------------------- | | Claude Code | `/plugin marketplace add olakai-ai/olakai-skills` | | Cursor | `npx add-skill olakai-ai/olakai-skills` | | VS Code Copilot | `npx add-skill olakai-ai/olakai-skills` | | Windsurf | `npx add-skill olakai-ai/olakai-skills` | | Cline | `npx add-skill olakai-ai/olakai-skills` | | Aider | `npx add-skill olakai-ai/olakai-skills` | NOTE: these are the available skills: - **olakai-new-project** - Build a new AI agent project from scratch with full observability - **olakai-integrate** - Add Olakai to existing AI code with minimal changes - **olakai-troubleshoot** - Diagnose and fix issues with events, KPIs, or SDK integration - **olakai-reports** - Generate usage summaries, KPI trends, ROI reports from the terminal ### 4) Authenticate or Create an Account Do you already have an Olakai account?

Yes Ask your coding agent to get you in (recommended): ``` /olakai-get-started ``` Alternatively, you may also use the CLI to verify: ```bash olakai login # Opens browser olakai whoami # Verify you're connected ```

No Ask your coding agent to set you up and follow the account creation steps (recommended) ``` /olakai-get-started ``` Alternatively, you may also sign up directly at **[app.olakai.ai/signup](https://app.olakai.ai/signup)**, then install the CLI and log in: ```bash npm install -g olakai-cli # Installs the CLI olakai login # Opens browser olakai whoami # Verify you're connected ```

### 5) Ask your coding agent to do the work With skills installed, your AI coding assistant understands Olakai. Open Claude Code/Copilot, etc and just describe what you need: ```bash I want to register the agent we implemented on this project with Olakai. Analyze the agent code first and help me brainstorm which KPIs would be helpful. Make sure to consider what data needs to be passed via the Olakai's customData variable. Then propose a plan to deploy Olakai. ``` Here are other examples of prompts you can use: | What you want | What to say | | ---------------------------------- | -------------------------------------------------------------------------------- | | Enable monitoring on existing code | "Enable Olakai monitoring on my agent" | | Choose the right KPIs | "What KPIs do you recommend for this agent, and how do I set them up in Olakai?" | | Fix missing data | "My events aren't appearing in Olakai, help me troubleshoot" | The skill handles the full workflow — agent registration, SDK integration, custom data setup, KPI configuration, and validation — all within your coding agent's chat. Here's a typical deployment plan your coding agent will generate: ![Typical deployment plan generated by an AI coding agent](/doc-images/typical-plan.png) ### 6) Verify data is flowing It's important to check that the data is flowing to Olakai. If you used the skill in the previous step, validation is usually handled automatically as part of the deployment plan: ![Deployment complete with validation](/doc-images/deployment-complete.png) If you need to verify manually, you can use these commands: ```bash Run the agent (as a test) and verify the data is captured by Olakai by listing the last few activities. Here are some example commands you can use: olakai activity list --limit 1 --json olakai activity get EVENT_ID --json | jq '{customData, kpiData}' Check that `customData` contains your fields, and `kpiData` shows numbers (not strings or null). ``` ### 7) See your traffic in the Live Events Feed Go to [Developers > Live Events Feed](https://app.olakai.ai/dashboard/developers/overview#events) to see your agent's traffic flowing through the platform in real time: ![Live Events Feed showing processed agent events](/doc-images/live-events-feed.png) Click on any event to open the Event Details view, which shows the full raw JSON payload alongside automated Olakai Checks — payload inspection for KPIs, configuration inspection, and payload hygiene validation for every custom data field: ![Event Details with raw JSON and Olakai Checks](/doc-images/event-details.png) The [Live Events Feed](/docs/olakai/developers-live-events) is a real-time stream of every event ingested into your account and your primary debugging tool. Use it to: - **Validate data is received by Olakai** — send a test event, then confirm it appears with a green "Processed" checkmark and a passing Olakai Check. - **Debug KPI values** — click any event to open the Event Details drawer. Automated checks (Payload Inspection, Configuration Inspection, Payload Hygiene) show which variables are present, missing, or have type mismatches. - **Diagnose Custom Data issues** — Configuration Inspection flags fields your SDK sends that aren't registered as Custom Data Configs. Unregistered fields are silently ignored. - **Spot type mismatches** — Payload Hygiene highlights fields where the SDK sends a string but Olakai expects a number (or vice versa). This is the most common cause of unexpected KPI values. - **Filter by new agents** — toggle "New Agents/Workflows" to see only events from agents created in the last 48 hours. Each event's details drawer includes a coding agent suggestion card with copy-paste guidance for fixing detected issues. ### 8) [You're live! Explore your dashboards](#youre-live-explore-your-dashboards) ## CLI Reference For the full CLI command reference — including agents, workflows, KPIs, custom data, activity, and environment variables — see the [Olakai CLI documentation](/docs/olakai/olakai-cli). ## Troubleshooting For visual debugging, open the [Live Events Feed](/docs/olakai/developers-live-events) at **Developers > Live Events Feed**. Click any event to inspect its raw payload, see which KPI variables are present or missing, and check for configuration or type mismatch issues. For CLI-specific troubleshooting (authentication errors, wrong environment, timeouts), see the [CLI Troubleshooting section](/docs/olakai/olakai-cli#troubleshooting). For agent integration troubleshooting (missing events, wrong KPI values, custom data issues), use the `olakai-troubleshoot` skill or see the [troubleshooting documentation](/docs/olakai/agent-troubleshooting). ## You're live! Explore your dashboards Your integration is complete. Log in to [app.olakai.ai](https://app.olakai.ai) and explore what Olakai is already tracking for you: - **Agent IQ** — ROI, governance compliance, and performance across your AI agents and workflows. - **Assistive IQ** — Adoption, usage, and efficiency metrics for chat and copilot tools. - **Developers** — Overview of your configuration and the Live Events Feed for real-time debugging. Hand off to your analyst or business stakeholder — they can start using dashboards immediately without any additional setup.