> ## Documentation Index
> Fetch the complete documentation index at: https://docs.costhawk.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenAI Integration

> Use OpenAI Admin API sync for visibility, and wrapped-key proxy routing for runtime controls

OpenAI supports two CostHawk integration modes.

## Mode A: Admin API Sync (Recommended start)

Use an OpenAI organization admin key for low-friction reporting.

* Endpoint coverage:
  * `/v1/organization/usage/completions`
  * `/v1/organization/costs`
* No app code changes required
* Best for quick organization-wide visibility

<Card title="Admin API Setup" icon="key" href="/integrations/admin-api">
  Use this first for reporting coverage.
</Card>

## Mode B: Proxy + Wrapped Key (Advanced controls)

Use proxy routing when you need strict runtime controls and enforcement.

<Warning>
  A wrapped key only works against the CostHawk proxy. Do not send `ch_sk_...` directly to `api.openai.com`.
</Warning>

<Info>
  If you are not sure whether you have a wrapped key or a CostHawk access token, read <a href="/integrations/wrapped-keys">Wrapped Keys</a> first.
</Info>

### Proxy URL

```
POST https://costhawk.ai/api/proxy/openai
```

This replaces `https://api.openai.com/v1/chat/completions`.

### Python (OpenAI SDK)

```python theme={null}
from openai import OpenAI

client = OpenAI(
    api_key="ch_sk_your_wrapped_key_here",
    base_url="https://costhawk.ai/api/proxy/openai",
)

response = client.chat.completions.create(
    model="gpt-4.1",
    messages=[{"role": "user", "content": "Hello, world!"}]
)

print(response.choices[0].message.content)
```

### Node.js (OpenAI SDK)

```javascript theme={null}
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "ch_sk_your_wrapped_key_here",
  baseURL: "https://costhawk.ai/api/proxy/openai",
});

const response = await client.chat.completions.create({
  model: "gpt-4.1",
  messages: [{ role: "user", content: "Hello, world!" }],
});

console.log(response.choices[0].message.content);
```

### curl

```bash theme={null}
curl -X POST https://costhawk.ai/api/proxy/openai \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ch_sk_your_wrapped_key_here" \
  -d '{
    "model": "gpt-4.1",
    "messages": [{"role": "user", "content": "Hello, world!"}]
  }'
```

## Recommended Pattern

1. Start with Admin API sync for baseline visibility.
2. Keep MCP enabled for assistant-native analysis and operations.
3. Add proxy routing for workloads that need policy enforcement or hard-stop limits.