Ir para o conteúdo
AI.info

AI Tools

Portkey Guardrails

Portkey Guardrails: preços, capacidades, atualizações e alternativas, lidos nas páginas do próprio desenvolvedor.

O site de Portkey Guardrails

Em inglês

Key facts

Price
Not stated on the maker’s pages
Free plan
Yes Read on portkey.ai
Platforms
Web · API
API
Yes
Open source
Not stated on the maker’s pages
Last update
29 Apr 2026 Read on new.portkey.ai

Portkey Guardrails adds checks to AI request pipelines. It supports rules for formats and content, partner guardrail services, and custom checks connected by webhook. Teams can apply checks to inputs, outputs, and embedding requests, and review verdicts in logs.

The hosted Free Forever plan lists deterministic guardrails; the Production plan lists LLM and partner guardrails, and Enterprise lists custom hooks and advanced evaluation templates. Portkey is a gateway and guardrail layer, not a model-training service.

Category
Developer Tools
Read from the maker’s pages
2026-11-18

Company

Portkey

Features

  • Check inputs for prompt injections and outputs for PII or unreliable responses
  • Apply deterministic checks such as JSON validation and regular expressions
  • Connect custom guardrails through webhooks
  • Run guardrails on embedding requests
  • Deny risky requests, retry failures, or route requests to another model
  • View guardrail verdicts, latency, and pass/fail status in logs
  • Set organization-wide guardrail policies across teams, workspaces, and models

Use cases

  • Block suspicious prompts before they reach a language model
  • Check AI responses for sensitive information before showing them to users
  • Validate structured responses before passing them to an application
  • Route or retry requests when a guardrail check fails
  • Filter embedding requests to protect sensitive data

Pricing

Free Forever

Free

  • 10k recorded logs per month
  • Deterministic Guardrails
  • Community Support

Production

$49/month

  • 100k recorded logs per month
  • LLM & Partner Guardrails
  • Role-Based Access Control
  • Production Support

Enterprise

Custom Pricing

  • Custom Guardrail Hooks
  • Advanced Evaluation Templates
  • Custom retention periods
  • SSO and granular budget and rate limits

Capabilities

  • Self-hosted — “Host it Yourself” source
  • No code retained — “Privacy Mode” source
  • API — “all through a single unified API.” source
  • Traces and evaluates — “View guardrails verdicts, latency, and pass/fail status for every check in real time.” source

Latest updates

  • New Integration: Hermes Agent + Portkey

    Point Hermes at Portkey's endpoint for 3,500+ models, request logging, cost tracking per session, budget limits, and MCP governance.

  • Agent Gateway - now live on Beta 🚀

    Register any agent, swap in your Portkey endpoint, and get governance, access control, observability, fallbacks, load balancing, and guardrails.

  • Pi Agent is now on Portkey

    Run Pi through Portkey for observability across usage, latency, and spend; budget and token limits; fallbacks, retries, and caching across 3000+ models; governance and…

  • New: Portkey × Conductor integration

    Use Portkey with Conductor to access 1600+ models, add fallbacks, load balancing, caching and routing, and get logs and metrics.

  • Secret References

    Use secret_mappings to fetch credentials at runtime from AWS Secrets Manager, Azure Key Vault, or HashiCorp Vault.

Where it works

Web · API

Alternatives to Portkey Guardrails

  • PRISMA AIRS AI Gateway

    Routes and governs AI traffic between applications, agents and model providers, with controls for security, usage and observability.

  • Helicone

    Helicone routes, logs, and analyzes AI application requests, helping teams debug model calls, track costs, and manage prompts.

  • Cyera AI Guardian

    Cyera Agent Guardian inventories AI agents, maps their data access, and helps security teams govern and block risky actions.

  • Dataiku LLM Mesh

    Dataiku LLM Mesh routes and governs access to hosted and self-hosted language models, with controls for usage, cost, performance, and safety.

  • LiteLLM

    LiteLLM routes requests to multiple LLM providers through one OpenAI-compatible API, with a Python SDK and a self-hosted gateway.

  • Langfuse

    Langfuse traces AI applications, manages prompts, and evaluates model and agent outputs using cloud or self-hosted deployments.

More from Portkey

  • PRISMA AIRS AI Gateway

    Routes and governs AI traffic between applications, agents and model providers, with controls for security, usage and observability.

Questions about Portkey Guardrails

Is Portkey Guardrails free?

Yes. The maker’s pages state that Portkey Guardrails can be used for free. Read on the maker’s pages on 18 Nov 2026.

Does Portkey Guardrails have an API?

Yes. The maker’s pages state that Portkey Guardrails has an API. Read on the maker’s pages on 18 Nov 2026.

Which platforms does Portkey Guardrails run on?

Portkey Guardrails runs on Web and API, as its maker’s pages state. Read on the maker’s pages on 18 Nov 2026.

Is Portkey Guardrails still available?

Yes, as of 18 Nov 2026: its maker’s pages were read that day and carried no notice that Portkey Guardrails has ended.

What changed recently in Portkey Guardrails?

The latest entry in its maker’s changelog: 29 Apr 2026, “New Integration: Hermes Agent + Portkey”. Before it: 21 Apr 2026, “Agent Gateway - now live on Beta 🚀”. Changelog checked on 18 Nov 2026.

Sources

Not confirmed

Open source
The checked pages describe the Gateway and guardrail logic as open source but do not identify a license.
Latest version
The product pages do not list a Portkey Guardrails version or model name.
Official website