AgentWatch

Appearance

Quickstart ​

Get AgentWatch running in under 2 minutes. This guide covers installation, configuration, and your first budget-protected API call.

Zero Dependency ​

AgentWatch requires no proprietary SDK. You use the standard OpenAI or Anthropic SDKs exactly as you normally would.

Instead of an SDK, you configure your client to point to the AgentWatch Edge Proxy and provide a combined API key.

Step 1: Install the Standard SDK ​

If you haven't already, install the official SDK for your provider.

Python ​

bash
pip install openai

TypeScript ​

bash
npm install openai

Step 2: Initialize with Combined Key ​

AgentWatch uses a "Bring Your Own Key" (BYOK) architecture. You combine your AgentWatch Developer Token with your real Provider API Key using a colon (:) separator.

aw_live_xxx:sk-proj-yyy

AgentWatch parses this key, uses the left half for identity and budget enforcement, and forwards the right half securely to the provider.

Python ​

python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.agent-watch.dev/v1/proxy/openai",
    api_key="aw_live_your_token:sk-proj-your-real-key"
)

TypeScript ​

typescript
import OpenAI from 'openai';

const client = new OpenAI({
    baseURL: "https://api.agent-watch.dev/v1/proxy/openai",
    apiKey: "aw_live_your_token:sk-proj-your-real-key"
});

Step 3: Make API Calls ​

Use the client exactly as you normally would. Budget enforcement happens automatically at the edge.

python
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Refactor this module..."}]
)
print(response.choices[0].message.content)

If the session exceeds its budget limit, the API will return an HTTP 402 Payment Required error, instantly blocking the execution before the upstream LLM provider is billed.

What Just Happened ​

  1. Your request was securely routed through the AgentWatch edge proxy.
  2. A sub-millisecond pre-flight budget check was performed against Cloudflare KV.
  3. Token costs were calculated in real-time using the provider's pricing.
  4. Telemetry was logged asynchronously (adding zero latency to your call).

Next Steps ​