Updated

Guardrails AI

Open-source Python framework that wraps LLM calls with validators from the Guardrails Hub (PII, toxicity, hallucination, format checks) to block or fix bad inputs and outputs; the company, also maker of the Snowglobe simulator, was acquired by Harvey in September 2026.

3 open-source projects declare Guardrails AI in their code; no maker's live product we track uses it yet.

Open sourceSelf-hostableFree tier

Acquired · Sep 2026: Harvey acquired Guardrails AI and its team joined Harvey's product and engineering organisation. The announcement says nothing about the open-source framework's future; the repo was still receiving commits in October 2026. Source ↗

Profile checked

Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

How makers use Guardrails AI

What Guardrails AI is used for

AI Guardrails & Security · Open-source library →Wrapping LLM calls with validators from the Guardrails Hub for PII, toxicity, hallucination and output format.Trade-off: Harvey acquired the company in September 2026 and said nothing about the framework's future, though the repo is still active.
Lakera1Nightfall1Prompt Security1NeMo Guardrails0Guardrails AI0Makers' products on each, for AI guardrails & security

When to use Guardrails AI, and when not

Pick Guardrails AI if

Something else fits better if

  • You want an API to screen prompts and outputs for injection and PII, without maintaining detectors: Lakera or Pangea.
  • You want to test your app for prompt injection and data leaks before release, in CI: Promptfoo.
  • You already monitor your models and want guardrails and evals in the same place: Arthur.
From each decision's guide, written from what makers' products chose and why.

Who uses it

No maker's live product on record yet, and 3 open-source projects that declare it in their code.

We haven't found a maker's live product that uses Guardrails AI yet — only the open-source projects below, which declare it in their code.

Opik
OpikOpen-source LLM tracing, evaluation and prompt optimization platform from Comet, self-ho…AI Guardrails & SecurityIn its code · source ↗

Open source: a project that declares Guardrails AI as a dependency in its public code — verifiable, but not necessarily a live product.

Alternatives to Guardrails AI

All alternatives by situation →

Questions makers ask about Guardrails AI

When should you use Guardrails AI?

Your app is in Python and you want checks on inputs and outputs that run in your own code (AI guardrails & security). Something else fits better if: you want an API to screen prompts and outputs for injection and PII, without maintaining detectors, pick Lakera or Pangea; you want to test your app for prompt injection and data leaks before release, in CI, pick Promptfoo.

Is Guardrails AI free?

Yes — it is free, open-source software.

Is Guardrails AI open source or self-hostable?

Open source, and you can self-host it. source ↗

Can AI coding agents work with Guardrails AI?

No llms.txt, official MCP server or CLI found yet.

Who uses Guardrails AI?

No maker's live product we track yet; 3 open-source projects declare it in their code. source ↗