Home
Documentation
AI Gateway
Policies

Policies

Apply guardrails, prompt enrichment, and rate limiting to a route.

The Policies module defines what AI Gateway does with each request that reaches a route before forwarding it to the LLM Provider, and what it does with the returned response. Each policy is linked to a single route and combines three optional blocks: Prompt Guard (filtering by moderation or regex), Prompt Enrichment (messages injected before the prompt), and Rate Limit (quotas by requests or tokens).

IMPORTANT
You must configure at least one section to save the policy. The form lets you combine Prompt Guard (Request), Prompt Guard (Response), Prompt Enrichment, and Rate Limit in any arrangement, but at least one must be filled in.

Quick links

Learn how to:

Filter and list policies
Create a policy
Edit a policy
Delete a policy
Review the route field
Configure Prompt Guard (Request)
Configure Prompt Guard (Response)
Configure Prompt Enrichment
Configure Rate Limit

Listing

The Policies page lists registered policies with pagination. You can filter by:

Route (optional): route associated with the policy.

Use Search to apply the filter and Clear to reset it.

Columns shown in the table:

Policy ID: internal policy identifier.
Route: associated route.
Configured sections: indicates which sections are configured (Prompt Guard / Prompt Enrichment / Rate Limit).
Created At: creation date and time.
Actions: Edit and Delete icons for each row.

Create Policy

Step 1

On the Policies page, click Create Policy.

Step 2

Select the target Route. This field is a searchable select. Type part of the endpoint to filter.

Step 3

Fill in at least one of the four optional sections: Prompt Guard (Request), Prompt Guard (Response), Prompt Enrichment, or Rate Limit. Each section has its own required-field rules. See the subsections below.

Step 4

Click Save to confirm the registration. To discard the operation, click Cancel.

Edit Policy

In the Actions column of the row you want, click the Edit icon. The Update Policy form opens with the current data. Change the required fields and click Save.

Delete Policy

In the Actions column of the row you want, click the Delete icon. Confirm the operation in the "Do you want to delete policy [policyId]?" dialog.

IMPORTANT
Deletion is permanent and cannot be undone. Deleting a policy makes the associated route accept requests without guardrails, prompt enrichment, or rate limiting.

Required field

Field	Type	Required	Rules
Route	Searchable select	Yes	Lists registered routes.

Prompt Guard (Request)

This optional block validates the request before it reaches the LLM provider. It combines three subfeatures: moderation, regex, and custom response.

MUTUAL EXCLUSION
Moderation and regex cannot coexist in the same Prompt Guard Request. Choose one path. When you fill in any moderation field, the regex fields become disabled, and vice versa.

Fields: Moderation subblock

Field	Type	Required	Rules
Moderation model	Select	Conditional: required when Moderation API Key is provided	Only value currently available: omni-moderation-latest (OpenAI moderation model).
Moderation API Key	Select	Conditional: required when Moderation model is provided	Lists registered API Keys, displayed as `<apiKeyId> (<provider>)`.

Fields: Regex subblock

Field	Type	Required	Rules
Regex action	Select	Conditional: required when Regex patterns have at least one pattern	Values: Reject, Mask.
Regex patterns	List of tags	Conditional: required when Regex action is provided	List of regular expressions.

Fields: Custom response subblock

Custom response returned when the request is rejected by moderation or regex.

Field	Type	Required	Rules
Custom response message	Text	Conditional: required when Custom response status code is provided	Minimum 1 character.
Custom response status code	Integer	Conditional: required when Custom response message is provided	Between 200 and 599.

Prompt Guard (Response)

This optional block validates the response returned by the provider before returning it to the requester. There is no moderation here, only regex and custom response, with the same pairing logic as the request.

STREAMING RESPONSES
For now, only responses returned without streaming are masked by the expressions configured in this section.

Fields: Regex subblock (response)

Field	Type	Required	Rules
Response regex action	Select	Conditional: required when Response regex patterns has at least one pattern	Values: Reject, Mask.
Response regex patterns	List of tags	Conditional: required when Response regex action is provided	List of regular expressions applied to the returned content.

Fields: Custom response subblock (response)

Field	Type	Required	Rules
Response custom message	Text	Conditional: required when Response custom status code is provided	Minimum 1 character.
Response custom status code	Integer	Conditional: required when Response custom message is provided	Between 200 and 599.

Prompt Enrichment

This block is optional as a section, but has required internal rules when enabled. When you add at least one message, you must provide at least one complete entry (role + content). Messages are prepended to the original prompt in the order they appear in the list.

Use Add message to add a new Message entry and the delete icon to remove entries.

Fields by message

Field	Type	Required	Rules
Role	Select	Yes (per message)	Values: system, assistant, user, function, tool, developer.
Content	Long text (textarea)	Yes (per message)	Minimum 1 character.

Rate Limit

This optional block limits consumption by time window. When enabled, it requires you to choose exactly one limit type: requests or tokens.

MUTUAL EXCLUSION
The policy cannot include requestCount and tokenCount at the same time. Choose one of them.

The screen shows a Count by selector with Token and Request buttons that determine which format is used.

Fields

Field	Type	Required	Rules
Rate limit type (Count by)	Toggle (Token / Request buttons)	Conditional: required when Rate Limit is enabled	Choose Token or Request. Mutually exclusive.
Rate limit value	Integer	Conditional: required when a Rate limit type is selected	Positive integer.
Time unit	Select	Conditional: required when Rate Limit is enabled	Values: Seconds, Minutes, Hours.

How happy are you with this page?

Routes

MCP Servers