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

# Integration with Cloudflare

> Track page visits at the edge by deploying a Cloudflare Worker that forwards page_visit events to The Prompting Company

The Cloudflare Worker collector tracks page visits **at the edge**,
before requests reach your origin. The worker sits in front of your existing site,
transparently proxies every matching request, and forwards page visit events to The Prompting Company — without
changing your origin server or adding any client-side JavaScript.

This guide will walk through how to set up the Cloudflare Worker with your deployment.

## How it works

The worker attaches to one or more routes and acts as a pass-through proxy:

1. A visitor requests a page that matches a configured route (for example, `example.com/*`).
2. The worker forwards the original request to your origin and returns the response immediately.
3. In parallel, the worker sends a page visit event to The Prompting Company. This happens in the background, so it never delays the response to the visitor.

## Before you start

You will need:

* A **Cloudflare account** with the domain you want to track added as a zone.
* A **TPC API key** with the `analytics:write` scope. You can create an API key with the `analytics:write` scope going to your Organization Settings, then selecting API Keys. Note that creating an API key requires an account with the Enterprise Plan.
* The `cf-collector` folder from the [samples repository](https://github.com/promptingcompany/samples/)

## Step 1: Install dependencies

From the `cf-collector` directory, install the dependencies:

```bash theme={null}
npm ci
```

## Step 2: Set your API key

The API key is stored as an encrypted [Wrangler secret](https://developers.cloudflare.com/workers/configuration/secrets/), not in the config file. Set it by running:

```bash theme={null}
npx wrangler secret put TPC_API_KEY
```

Paste your key when prompted.

<Warning>
  Never commit your API key or place it in `wrangler.jsonc`. Always store it as a secret.
</Warning>

## Step 3: Configure your routes

Edit the `routes` array in `wrangler.jsonc` to match the pages you want to track. Set the `pattern` to your domain and the `zone_name` to the matching Cloudflare zone:

```jsonc wrangler.jsonc theme={null}
{
  "routes": [
    {
      "pattern": "example.com/*",
      "zone_name": "example.com"
    }
  ]
}
```

You can add more entries to the array to track additional domains or path patterns.

## Step 4: Deploy

To deploy the worker to Cloudflare, run:

```bash theme={null}
npm run wrangler:deploy
```

Once deployed, Cloudflare routes all traffic matching your patterns through the worker, and page visits should begin appearing in your analytics page.

## Troubleshooting

### Events are not appearing

* Confirm the request URL matches a `pattern` in your `routes` array, including the correct `zone_name`.
* Make sure the `TPC_API_KEY` secret is set for the environment you deployed to (`npx wrangler secret list`).
* Verify the worker is deployed and bound to the route in the Cloudflare dashboard under **Workers & Pages**.

### Inspect live logs

Stream logs from the deployed worker to see event delivery and any errors in real time:

```bash theme={null}
npm run tail
```