Each Guardrail Check has a specific purpose, it’s own parameters, supported hooks, and sources.

​
Partner Guardrails

acuvity-logo

Acuvity

  • Scan Prompts
  • Scan Responses For PII, toxicity, prompt injection detection, and more.
logo-1

Apora

  • Validate Aporia policies
  • Define your Aporia policies on your Aporia dashboard and just pass the project ID in Portkey Guardrail check.
microsoft-assure

Azure

  • Detect and Redact PII
  • Apply Azure’s Content Safety checks
bedrock-logo

AWS Bedrock

  • Analyze and redact PII to avoid model manipulation
  • Bring your AWS Guardrails directly inside Portkey and more!
lasso

Lasso Security

  • Scan Content Lasso Security’s Deputies analyze content for various security risks including jailbreak attempts, custom policy violations, and more.
mistral

Mistral

  • Detect and filter harmful content across multiple dimensions
logo-4

Pangea

  • Text Guard for scanning LLM inputs and outputs
  • Analyze and redact text to avoid model manipulation
  • Detect malicious content and undesirable data transfers
logo-3

Patronus

  • Hallucination detection
  • Check for conciseness, helpfulness, politeness
  • Check for gender, racial bias
  • and more!
logo-2

Pillar

  • Scan Prompts
  • Scan Responses For PII, toxicity, prompt injection detection, and more.

The logic for all of the Guardrail Checks (including Partner Guardrails) is open source.

View it here and here on the Portkey Gateway repo.

​
Bring Your Own Guardrail

We have built Guardrails in a very modular way, and support bringing your own Guardrail using a custom webhook! Learn more here.

​
Portkey’s Guardrails

Along with the partner Guardrails, there are also deterministic as well as LLM-based Guardrails supported natively by Portkey.

BASIC Guardrails are available on all Portkey plans.

PRO Guardrails are available on Portkey Pro & Enterprise plans.

​
BASIC — Deterministic Guardrails

Regex Match


Checks if the request or response text matches a regex pattern.

Parameters: rule: string
Supported On: input_guardrails, output_guardrails

Sentence Count


Checks if the content contains a certain number of sentences. Ranges allowed.

Parameters: minSentences: number, maxSentences: number
Supported On: input_guardrails, output_guardrails

Word Count


Checks if the content contains a certain number of words. Ranges allowed.

Parameters: minWords: number, maxWords: number
Supported On: input_guardrails, output_guardrails

Character Count


Checks if the content contains a certain number of characters. Ranges allowed.

Parameters: minCharacters: number, maxCharacters: number
Supported On: input_guardrails, output_guardrails

JSON Schema


Check if the response JSON matches a JSON schema.

Parameters: schema: json
Supported On: output_guardrails only

JSON Keys


Check if the response JSON contains any, all or none of the mentioned keys.

Parameters: keys: array, operator: string
Supported On: output_guardrails only

Contains


Checks if the content contains any, all or none of the words or phrases.

Parameters: words: array, operator: string
Supported On: output_guardrails only

Valid URLs


Checks if all the URLs mentioned in the content are valid

Parameters: onlyDNS: boolean
Supported On: output_guardrails only

Contains Code


Checks if the content contains code of format SQL, Python, TypeScript, etc.

Parameters: format: string
Supported On: output_guardrails only

Lowercase Detection


Check if the given string is lowercase or not.

Parameters: format: string
Supported On: input_guardrails, output_guardrails

Ends With


Check if the content ends with a specified string.

Parameters: Suffix: string
Supported On: input_guardrails, output_guardrails

Webhook


Makes a webhook request for custom guardrails

Parameters: webhookURL: string, headers: json
Supported On: input_guardrails, output_guardrails

​
PRO — LLM Guardrails

Moderate Content


Checks if the content passes the mentioned content moderation checks.

Parameters: categories: array
Supported On: input_guardrails only

Check Language


Checks if the response content is in the mentioned language.

Parameters: language: string
Supported On: input_guardrails only

Detect PII


Detects Personally Identifiable Information (PII) in the content.

Parameters: categories: array
Supported On: input_guardrails, output_guardrails

Detect Gibberish


Detects if the content is gibberish.

Parameters: boolean
Supported On: input_guardrails, output_guardrails

You can now have configurable timeouts for Partner & Pro Guardrails!

​
Contribute Your Guardrail

Integrate your Guardrail platform with Portkey Gateway and reach our growing user base. Check out some existing integrations to get started.