gateway

• Run the development `server npm run start``
• Deploy your application `npm run deploy``
• Read the documentation https://developers.cloudflare.com/workers
• Stuck? Join us at https://discord.gg/cloudflaredev

Configuration

Header	Description
x-gateway-service-host	The host for the API request
x-gateway-service-token	The API token to use if not in the environment variables
x-gateway-service-auth-key	The authorization key for the request. This is used in combined with x-gateway-service-auth-type. If x-gateway-service-auth-type is HEADER then this value will be used as the HTTP header; if x-gateway-service-auth-type then this value will be the query field.
x-gateway-service-auth-type	The authorization type. This is either HEADER to add the token to the requests header or QUERY to add it to the requests parameters.
x-gateway-service-auth-prefix	Any prefix for the token. A space is appended after this prefix.
x-gateway-service-type	Indicates whether the downstream service is GATEWAY or DIRECT. The default is DIRECT.
x-gateway-service-proxy	Required only when x-gateway-service-type is GATEWAY. The format can be anything, provided the environment has a secret named {x-gateway-service-host}_API_KEY_{x-gateway-service-proxy}. For example, if the GATEWAY is CLOUDFLARE AI and the service is MISTRAL, then x-gateway-service-proxy can be MISTRAL. Note that this proxy header will be used to resolve the Cloudflare secret in the environment by replacing all non-alphanumeric characters with an underscore (_).

Example

• If you want to use Bearer/token authentication then you will need to set the following headers appropriately:

Header	Value
x-gateway-service-auth-key	Authorization
x-gateway-service-auth-type	HEADER
x-gateway-service-auth-prefix	Bearer

• If you want to use a gateway service where the host remains constant but different API keys are used depending on the specific service within the gateway, follow these steps. For example, if you're using the Cloudflare AI gateway with the OpenAI service, and you have the Cloudflare secret for this service set as GATEWAY_AI_CLOUDFLARE_COM_API_KEY_OPEN_AI = sk-your-api-key, then make a request with the following headers:

Header	Value
x-gateway-service-auth-key	Authorization
x-gateway-service-auth-type	HEADER
x-gateway-service-auth-prefix	Bearer
x-gateway-service-type	GATEWAY
x-gateway-service-host	gateway.ai.cloudflare.com
x-gateway-service-proxy	OPEN AI

POST https://gateway.your-subdomain.workers.dev/v1/{account_id}/{gateway_id}/openai/chat/completions
x-gateway-service-auth-key: Authorization
x-gateway-service-auth-type: HEADER
x-gateway-service-auth-prefix: Bearer
x-gateway-service-type: GATEWAY
x-gateway-service-proxy: OPEN AI
x-gateway-service-host: gateway.ai.cloudflare.com
Content-Type: application/json

{
  "model": "gpt-3.5-turbo",
  "messages": [
    {
      "role": "user",
      "content": "What is AI?"
    }
  ]
}

Errors

If configured incorrectly a HTTP 400 JSON response will be returned. The format will be { "error": "<error message>" }.

Logging

Metadata is collected for every request. This metadata is used to detect abuse and to help prevent unauthorized access.

All headers are recorded as a JSON blob. Known sensitive headers are stripped from the blob. In addition, the following request header fields are recorded separately:

Header	Description
user-agent	The user agent.
cf-connecting-ip	The client IP address connecting to Cloudflare to the origin web server.
cf-ipcountry	The two-character country code of the originating visitor’s country.
x-gateway-service-id	The requested service's identifier.
x-gateway-service-name	The requested service's human readable name.
x-gateway-identifier-for-vendor	An alphanumeric string that uniquely identifies a device to the connecting application’s vendor.
x-gateway-bundle-identifier	The connecting application's bundle identifier.
x-gateway-bundle-version	The connecting applications version number.

The following is also recorded:

• Full URL of the request
• Responses status code
• Any server specific errors

Data Retention

All metadata is kept for a configured amount of period. Beyond this timeframe, the metadata undergoes permanent deletion with no backup procedures in place.

To achieve this, a cron job executes every Sunday 12:00 AM(configurable), targeting metadata older than configured period for deletion. The configuration for this cron job is stored in the wrangler.toml file, specifically under the triggers section.

The cron job is designed to respect the DELETE_OLD_DATA_BEFORE environment variable. If this variable is not set, the cron job will intentionally generate an SQL error, preventing the deletion of any data.

DELETE_OLD_DATA_BEFORE variable Format

[sign] quantity unit

• sign: Optional. Use + for addition and - for subtraction.
• quantity: Numeric value.
• unit: Time unit (year(s), month(s), day(s), hour(s), minute(s), second(s)).