Getting Started
Getting Started
Get AgentWall running in 5 minutes.
Prerequisites
- An AgentWall API key (sign up here)
- An OpenAI API key (or other LLM provider)
- Python 3.8+ or Node.js 16+
Step 1: Get Your API Key
- Go to agentwall.io/admin
- Create an account or log in
- Navigate to API Keys → Create New Key
- Copy your key (starts with
aw-)
Step 2: Configure Your Client
Python (OpenAI SDK)
Node.js (OpenAI SDK)
cURL
Step 3: Make Your First Request
That's it! Your request is now protected by AgentWall.
Step 4: Enable Run Tracking (Recommended)
For multi-step agent tasks, use agentwall_run_id to track the entire run:
Step 5: View Your Dashboard
Go to agentwall.io/admin to see:
- Agent Runs - All your agent tasks with step-by-step details
- Request Logs - Every API request with costs and latency
- Budget Usage - Track spending against your limits
- Alerts - Loop detections, budget warnings, kills
What's Next?
- Concepts - Understand run tracking, loop detection, DLP
- API Reference - Full API documentation
- Examples - Real-world use cases
- LangChain Integration - Use with LangChain
Troubleshooting
"Invalid API Key" (401)
Make sure your API key:
- Starts with
aw- - Is copied correctly (no extra spaces)
- Is active in the dashboard
"Model not found" (404)
Check that:
- You're using a supported model (gpt-4, gpt-3.5-turbo, etc.)
- Your OpenAI API key is configured in AgentWall dashboard
High Latency
AgentWall adds <10ms overhead. If you see high latency:
- Check your network connection
- The LLM provider (OpenAI) may be slow
- Use streaming for better perceived performance
Need help? Contact support